Cancer

New data bring to a close Australia’s most intensive cancer cluster investigation Though only a small proportion of cancer clusters are reported in the peer-reviewed literature, a consistent picture emerges from those reports. The term “cancer cluster” refers to health authorities being alerted to a perceived increased incidence of cancer, involving 15 cases or fewer in the first instance, within a particular community or group. Aware of the anxiety generated by the prospect of a cancer cluster, cluster investigators focus on local circumstances that might account for increased risk, and if they fail to identify any such factors, the matter rests.1 The report by Sitas and colleagues in this issue 2 is a rare example of the implications of a cluster investigation being rigorously pursued. In this case, the implication was that a cluster among staff of the Australian Broadcasting Corporation (ABC) studios in Toowong, Brisbane, might reflect a higher-than-average risk of breast cancer among female employees of the ABC throughout Australia, due to some common cause.3 In terms of identifying causative agents, notification of an individual cluster can be likened to a case report. Nothing can be immediately proved; everything depends on a subsequent, more broadly based study. Cancer clusters, and their investigation, are distinguished from infectious disease clusters by, among other things, the longer time frame involved and the fact that cancer is the most feared disease in Australia and similar countries. Demands made in respect of cancer cluster investigations centre on whether the cluster is explicable by chance and, if not, what caused it. The Toowong cluster came to national attention via a television program entitled “One in a million”, an epithet derived from the cluster investigators’ initial assessment of the likelihood that the increased incidence of breast cancer might be attributable to chance.4 Calculation of the probability that a cluster is attributable to chance is limited, if not precluded, by a posteriori definition of the study population — often illustrated by reference to the Texas sharpshooter who was in the habit of firing at the side of a barn, locating the bullet hole and painting a target round it.5 But the community wants answers, not technicalities. In the case of Toowong, the difficulty inherent in making such a determination is illustrated by revision of the initial estimate, specified in the summary of the final report as follows: [T]he likelihood of this event occurring by chance is about one in a million. This, however, may oversimplify the situation and, when further analyses are performed, adjustment of the P-value for implied multiple comparisons increased its estimated value to 0.04. That potentially increases the likelihood of the cluster occurring by chance to one in 25.3 Whether cluster investigations have ever revealed a new carcinogen is debated. Though vinyl chloride and diethylstilboestrol are sometimes cited in this context, the carcinogenicity of these compounds was implicated from multiple case reports involving alert physicians rather than from the perceptions of people in the affected group. Indeed, diethylstilboestrol was initially investigated because a lift malfunction in Boston resulted in a gynaecologist and a pathologist exchanging experiences.6 Otherwise, increased numbers of cancer cases within communities located adjacent to point-source pollution may implicate particular carcinogens. However, assessments of lung cancer near steelworks or smelting operations, mesothelioma near asbestos mines, thyroid cancer near a failed reactor, or haematopoietic malignancy in the vicinity of toxic waste dumps7 are readily distinguished from cluster studies, the latter being prompted by increased incidence rather than any attempt to determine whether an increase exists in light of a specific exposure. Novel exposure to carcinogens may also be investigated following spatial aggregation studies, which typically involve hundreds or thousands of cases — for example, the investigation of exposure to pesticides or other agents in relation to increased breast cancer incidence on Long Island, New York.8 Once cluster investigations are distinguished from case reports, point-source pollution investigations and spatial aggregation studies, it is arguable that clusters have rarely, if ever, resulted in the discovery of a specific cause of cancer, much less revealed a new carcinogen. Accordingly, the result of the Toowong cluster investigation was typical. In Toowong, all plausible causes of the cluster, including radiofrequency and extremely low frequency electromagnetic fields, ionising radiation and chemical contamination of the site or its water supply, were excluded.3 The findings of Sitas and colleagues2 are a good outcome for women employed by the ABC, and for those who worked in Toowong specifically. The absence of increased risk of breast cancer among ABC employees Australia-wide, together with a failure to identify any agent that could account for increased risk in Toowong and the 1 in 25 probability that the situation may have arisen by chance,3 mean the case is closed. However frustrating it may be, chance emerges as the most likely explanation. ABC staff at Toowong and elsewhere have no reason to be apprehensive about being at increased risk. There are no reasonable medical or scientific grounds for such women to undergo more rigorous clinical examination or more frequent mammographic screening than is recommended for women in Australia generally. These considerations critically depend on the findings of Sitas et al.2 Likewise, any notion that the building or site at Toowong presents a toxic hazard is now little short of absurd. Hopefully, ABC staff will be so advised. Beyond that, however, this good news might be shouted from the top of Uluru for all the likelihood it has of reaching the wider community, whose perceptions about cancer causation now include the “one in a million” scenario. The media in Australia respond to an insatiable demand for news of unsuspected exposure to carcinogens. Overwhelmingly, reports about possible cancer causation refer to consumer products and food contamination.9 Apart from a correct appreciation of carcinogens in tobacco, perceived causes of cancer range from pesticide residues in food to deodorants, from artificial sweeteners to mobile phones.7 For each such agent, there exists evidence of potential harm — but the relevant risks are either vanishingly small or simply not established. Maintaining speculation about Toowong comes a recent report of another breast cancer case and related matters connected with Toowong (Box). In point of fact, more and more cases of breast cancer will occur among women who once worked at ABC Toowong, in the same way that more and more cases of breast cancer will occur among any cohort of Australian women initially identified as aged 30–50 years. Tragically, one Australian woman in nine will develop breast cancer by the age of 85 years.10 About 60 cases must be anticipated within the Toowong cohort of 550, taking account of no other risk factor than being a woman in Australia. The prospect, however, is that each such case will be reported as involving another Toowong employee, fuelling anxiety about a common cause of harm, with no mention being made of the work of Sitas et al.2 Reference to a lack of risk hardly contributes to the momentum of a great story. But there is a price. People who misunderstand cancer causation are less likely to engage in practices known to reduce the risk of disease.11 Any unwarranted anxiety about insidious exposure to carcinogens seems likely to impede adoption of evidence-based measures to prevent cancer. That, together with any needless burden of anxiety, is bad news. Headlines from the Brisbane Courier Mail, 22 and 26 February 2010

Bernard W Stewart PhD, FRACI, DipLaw

Objective: To determine whether there is an excess risk of breast cancer among female employees of the Australian Broadcasting Corporation (ABC), especially outside Queensland, compared with women in the general populations of the states and territories.Design, setting and participants: We used an occupational cohort design. Information from ABC staff records was linked with data from state and territory cancer registries to identify female employees of the ABC with an incident, histologically confirmed breast cancer. Data linkage was complemented by a self-report method. We included a cohort of ABC female employees who had developed breast cancer at any time between 1994 and 2005, during their employment or after cessation of employment with the ABC. The standardised incidence ratio (SIR) was calculated as the number of women at the ABC observed with breast cancer divided by the expected number based on population rates in each state and territory. Tests for heterogeneity were performed to examine the variation of breast cancer risk between states and territories.Results: Out of 5969 women who were permanently employed either part-time or full-time at the ABC between 1994 and 2005, 48 eligible women with breast cancer were identified. An excess risk of breast cancer among ABC female employees in Queensland (identified in an earlier study) was reconfirmed. No excess risk of breast cancer was observed among ABC staff diagnosed in states outside Queensland (SIR, 1.01 [95% CI, 0.72–1.38]), or in Australia as a whole (including Queensland) (SIR, 1.12 [95% CI, 0.83–1.49]). There was no significant heterogeneity in breast cancer risk among states and territories once Queensland was excluded from the analysis (P = 0.39).Conclusion: No statistically significant excess risk of breast cancer in ABC female employees was found across the Australian states and territories as a whole compared with their respective population incidences. A statistically significant increased risk of breast cancer was found among ABC female employees in Queensland, consistent with the findings in an earlier report.

Freddy Sitas MSc(Med), MSc(Epidemiol), DPhil · Dianne L O’Connell BMaths(Hons), PhD · Cathelijne H van Kemenade MSc, MPH · Mark W Short BSc(Hons), PhD · Kun Zhao BEcon, BPsych(Hons)

Incorrect formula and number of trial subjects in Box: In “It’s time to depolarise the unhelpful PSA-testing debate and put into practice lessons from the two major international screening trials” in the 5 April 2010 issue of the Journal (Med J Aust 2010; 192: 393-396), in the Box, there was an error in the total number of prostate cancers detected in the screening arm of the ERSPC trial and in the formula used to calculate differences between the two arms. The Box, with corrected figures in the last two columns and a replacement formula in a footnote, is reproduced here. Staging and prostate cancer deaths among men participating in the PLCO and ERSPC trials5,6 Control arm Screening arm STAGE OF CANCER AT DIAGNOSIS PLCO Trial n = 2940 n = 3437 RD (%)* Stage 1 15 (0.5%) 18 (0.5%) + 2.6% Stage 2 2790 (94.9%) 3297 (95.9%) + 1.1% Stage 3 56 (1.9%) 49 (1.4%) – 25.2% Stage 4 79 (2.7%) 73 (2.1%) – 21.0% ERSPC n = 4611 n = 6169 RD (%)* Stage T1a,b 277 (6.0%) 212 (3.4%) – 42.8% Stage T1c 1805 (39.1%) 3477 (56.4%) + 44.0% Stage T2 1320 (28.6%) 1755 (28.4%) – 0.6% Stage T3 749 (16.2%) 509 (8.3%) – 49.2% Stage T4 156 (3.4%) 67 (1.1%) – 67.9% Stage M1 304 (6.6%) 149 (2.4%) – 63.4% CAUSE OF DEATH (in men diagnosed with prostate cancer) PLCO Trial Rate ratio† Prostate cancer 82 92 1.11 Other causes‡ 225 312 — ERSPC Prostate cancer 326 214 0.80 Other causes ns ns ns ERSPC = European Randomized Study of Screening for Prostate Cancer. ns = not specified. PLCO = Prostate, Lung, Colon and Ovarian. RD = relative difference. * The relative difference is the percentage increase or decrease in cases of a given stage of prostate cancer detected by screening, using the formula RD = {[(proportion in screened arm ÷ proportion in control arm) × 100] – 100}%. eg, for the PLCO Trial, Stage 3, RD = {[(49/3437 ÷ 56/2940) × 100] – 100}% = –25.2%. † Rate ratio = adjusted value given by the authors.5,6 ‡ Men with prostate cancer dying of other causes.

James W Denham · Ron Bender · William E J Paradice

Two recently reported large-scale trials conducted in the United States and western Europe have provided evidence that coordinated screening programs will not reduce mortality in countries or regions where prostate-specific antigen (PSA) testing is already highly prevalent, but will reduce mortality in places where PSA testing prevalence is low. The trials also produce evidence that coordinated screening will cause over-diagnosis and over-treatment. The instigation of a national screening program should be delayed until a more specific marker for aggressive disease than PSA level becomes available. In the meantime, results of the two trials can be used to inform the development of regional testing policies in Australia. These policies should encourage regular PSA testing in regions with low testing prevalence, but must also embrace methods of dealing with over-diagnosis and over-treatment. “Active surveillance” programs (whereby men with early-stage cancers are monitored regularly by PSA testing and digital rectal examinations) and development of counselling services should be encouraged.

James W Denham MD, FRCR, FRANZCR · Ron Bender BASocSci, GradDipMgmt, MA · William E J Paradice BNatRes, MSc, PhD, FAICD

Objective: To describe the management and outcomes of a population-based cohort of patients with pancreatic cancer treated with chemotherapy or radiotherapy in Victoria, Australia.Design, setting and patients: Questionnaire-based study of patients diagnosed with pancreatic cancer during 2002–2003 in Victoria who were retrospectively identified from the Victorian Cancer Registry and followed up for a minimum of 5 years.Main outcome measures: Reported treatment, referral patterns and survival rates.Results: 1044 patients with pancreatic cancer were identified, of whom 927 were eligible for the study. Completed questionnaires were obtained for 831 eligible patients (response rate, 89.6%) and data for 66 patients with tumours of the ampulla of Vater and neuroendocrine tumours were excluded. Of the remaining 765 patients, 6.5% were managed in multimodality clinics. Chemotherapy was considered for 413 patients and radiotherapy was considered for 162. One-third of the cohort (275 patients) received chemotherapy, most commonly as palliative treatment (185). Single-agent gemcitabine was the most common palliative treatment (154), and was associated with a median overall survival of 6.6 months. Radiotherapy was used in 119 patients (15.6% of the cohort) — it was used alone or with chemotherapy, as postoperative adjuvant treatment, as potentially curative radical treatment, or as palliative treatment. For 45 patients with locally advanced disease who were treated with chemoradiation as radical treatment, median overall survival was 13.1 months.Conclusions: There appears to be under-referral of patients to medical and radiation oncologists. Median survival of patients treated with radical chemoradiation or palliative chemotherapy is consistent with clinical trial data, but outcomes for patients in our cohort were generally poor. Development and implementation of treatment guidelines may result in improved outcomes.

Michael Jefford MB BS, PhD, FRACP · Vicky Thursfield BSc, GradDipApplStats · Yvonne Torn-Broers BA(Hons) · Trevor Leong MB BS, MD, FRANZCR · Mario Guerrieri MB BS, FRANZCR · Tony Speer BE, MB BS, FRACP

To the Editor: I read with interest the article by Grange and colleagues suggesting that vaccination with BCG vaccine or past severe infections may help protect against the development of melanoma.1 They do not mention the work of Coley. Coley was a surgeon at the Hospital for the Ruptured and Crippled in New York, who, after observing the regression of tumours in patients who developed erysipelas involving the tumour site, reported in 1891 in the Annals of Surgery,2 and again in 1893 in the American Journal of the Medical Sciences,3 that injecting Streptococcus pyogenes from erysipelas isolates into the patient’s tumour induced regression. Thus the relationship between neoplasia and bacterial infection has been recognised for 118 years. Coley is considered the father of cancer vaccines.

Michael G E O’Rourke

To the Editor: We report a case of desmoplastic small round cell tumour (DSRCT) in the liver of a 23-year-old woman who presented with a recurrent non-pruritic rash. The woman had a generalised macular rash, predominantly on the back and upper thighs. She reported recurrent similar skin rashes over the previous 4 months that had been treated with intermittent courses of oral antibiotics. She had felt mild fatigue during the 3 weeks prior to presentation. She was otherwise well, but had non-tender hepatomegaly (17 cm) on examination. There was no peripheral lymphadenopathy. Levels of cholestatic liver enzymes were mildly raised (γ-glutamyl transpeptidase, 198 IU/L; alkaline phosphatase, 246 IU/L), but bilirubin and immunoglobulin levels were normal. The serum level of cancer antigen 125 (CA125) was 141 U/mL (reference range, < 35 U/mL). A computed tomography scan of the chest and abdomen showed multiple hepatic lesions (Box 1). A radiologically guided biopsy was taken, and histological examination revealed the typical morphology and immunophenotype of DSRCT (Box 2). Our patient was counselled regarding diagnosis and likely poor prognosis. We did not attempt to preserve fertility, because we felt treatment should not be delayed and that life expectancy was limited. Chemotherapy with alternating VAC (vincristine, doxorubicin and cyclophosphamide) and IE (ifosfamide and etoposide) cycles was started promptly. Molecular testing for EWS1/WT1 (see below) was not performed, as there was insufficient biopsy tissue available. The patient’s rashes disappeared after one treatment cycle, and we postulate that the rashes were paraneoplastic and immune-mediated. Restaging scans after four cycles of chemotherapy demonstrated a partial response. Stem cells were pre-emptively mobilised to store for possible subsequent autologous transplantation. The patient will be reassessed after eight cycles. If there is a significant response, the options of high-dose chemotherapy with autologous stem cell transplantation and/or debulking surgery will be explored. To our knowledge, this is the first Australian report of DSRCT in a woman presenting with recurrent rash. DSRCT is a rare, aggressive tumour that predominantly affects males in early adulthood.1 There is a single Australian report of a 15-year-old boy who died of DSRCT 20 months after diagnosis.2 The histogenesis of DSRCT is unknown, but it exhibits divergent differentiation, expressing epithelial, muscular and neural proteins. It is characterised by the chromosomal translocation t(11;22)(pl3;ql2), formed by fusion of the Ewing sarcoma gene (EWS1) to the Wilms tumour suppressor gene (WT1).3 Patients typically present with non-specific symptoms, and the tumours are usually intra-abdominal. The level of CA125 is often raised, but this does not assist with diagnosis or monitoring.1 Diagnosis is by histology and immunohistochemistry, complemented by cytogenetic identification of an EWS1/WT1 translocation. Patients with DSRCT have a poor prognosis, with a median survival of 15 months. Given the rarity of the tumour, there are no data from randomised phase III trials to guide management. Aggressive multimodality treatment offers the highest chance of disease control and prolonged overall survival.1,4 Palliative debulking can be of benefit if curative resection is not feasible. Radiotherapy is best employed as consolidation treatment after chemotherapy and surgery.1,5 Combination chemotherapy is the backbone of therapy and offers improved progression-free survival. The P6 protocol,6 which uses alternating cycles of VAC and IE with 21 days between cycles, is most widely employed. Subsequent high-dose myeloablative chemotherapy with autologous stem cell support may be beneficial.4 1 Pre-treatment computed tomography scan of the patient’s abdomen, showing multiple hypodense lesions in the enlarged liver 2 Histological sections from a core biopsy of a desmoplastic small round cell tumour (DSRCT) in the liver A: Irregular nests of small round hyperchromatic tumour cells were embedded in a prominent fibrotic stroma (haematoxylin–eosin stain, original magnification ×250). A diagnosis of DSRCT was confirmed by immunohistochemical stains (B–D). B: The muscle marker desmin. C: An epithelial marker AE1/3. D: The Wilms tumour marker WT1. (Original magnification of images B–D ×100.)

Meena Okera · David Moffat · Sudarshan Selva-Nayagam

Objective: To examine the symptoms, level of suffering, and care of Australian children with cancer at the end of life.Design, setting and participants: In a study conducted at the Royal Children’s Hospital, Melbourne, parents of children who had died of cancer over the period 1996–2004 were interviewed between February 2004 and August 2006. Parents also completed and returned self-report questionnaires.Main outcome measures: Proportions of children suffering from and treated for various symptoms; proportion of children receiving cancer-directed therapy at the end of life; proportion of children whose treatment of symptoms was successful; location of death.Results: Of 193 eligible families, 96 (50%) were interviewed. All interviews were conducted in person, and occurred a mean of 4.5 years (SD, 2.1 years) after the child’s death. Eighty-four per cent of parents reported that their child had suffered “a lot” or “a great deal” from at least one symptom in their last month of life — most commonly pain (46%), fatigue (43%) and poor appetite (30%). Children who received cancer-directed therapy during the end-of-life period (47%) suffered from a greater number of symptoms than those who did not receive treatment (P = 0.03), but the severity of symptoms did not differ between these groups. Of the children treated for specific symptoms, treatment was successful in 47% of those with pain, 18% of those with fatigue and 17% of those with poor appetite. Of the 61 families who felt they had time to plan where their child would die, 89% preferred to have their child die at home. The majority of children (61%) died at home. Of those who died in hospital, less than a quarter died in the intensive care unit.Conclusions: Relatively high rates of death at home and low rates of unsuccessful medical interventions suggest a realistic approach at the end of life for Australian children dying of cancer. However, many suffer from unresolved symptoms, and greater attention should be paid to palliative care for these children.

John A Heath PhD, FRACP · Naomi E Clarke BMedSc(Hons) · Susan M Donath BSc(Hons), PhD · Maria McCarthy BAppSc, MAppSc · Vicki A Anderson BA, PhD · Joanne Wolfe MD, MPH

To the Editor: The general practitioners’ dilemma is acute. Women asking for mammography and/or testing for ovarian cancer; men asking for prostate-specific antigen (PSA) testing for prostate cancer. The debate, fuelled by uncertain knowledge, rages in the general medical journals,1,2 including contributions in the 17 August 2009 issue of the MJA.3,4 The PSA test is, of course, intimately associated with the name of its inventor, Thomas A Stamey. I am, therefore, somewhat surprised that neither of these recent MJA articles mentioned Stamey’s 2004 “recanting” of his proposed value of the test back in 1987: “current evidence from the last 10 years is convincing that the relationship between prostate cancer and serum PSA is tenuous at best, especially with serum PSA less than 10 ng/mL and perhaps even less than 22 ng/mL. This time is not the first we have had second thoughts regarding the usefulness of serum PSA in preoperatively reflecting prostate cancer”. He concludes: “any excuse to biopsy the prostate has an excellent, age dependent chance of being positive”.5 Perhaps the authors could tell your readers why they did not think that Stamey’s 20-year experience of PSA testing and his 2004 conclusions warranted a mention?

Peter C Arnold

In reply: The letter by Arnold highlights the practical dilemma facing busy doctors when communicating with patients about whether to screen for various cancers, prostate cancer in particular. There are no easy answers at present, although presentation of both sides of the argument regarding prostate cancer testing helps doctors and patients make a shared decision.1 We believe the following decision-aid show card, written in plain English, is an excellent resource to assist with this: <http://www.cancer.org.au/File/HealthProfessionals/GPprostateshowcard.pdf >.

David P Smith · Emily Banks · Mark S Clements · Robert A Gardiner · Bruce K Armstrong

A history of the ‘unfortunate experiment’ at National Women’s Hospital. Linda Bryder. Auckland: Auckland University Press, 2009 (vi + 250 pp). ISBN 978 1 86940 435 2. The so-called “unfortunate experiment” of this book triggered the 1987 New Zealand Cartwright Inquiry into cervical cancer treatment. Consultant gynaecologist Herbert Green of the Auckland National Women’s Hospital was the main target of a long and complex legal process hinging on allegations that he had prospectively studied two groups of women referred to the hospital with positive cervical cytology. One large group received treatment, following which most had no further problems (although some later developed invasive cancer). The second group, it was said, had no treatment but, despite positive cytology, were followed for years so that the natural history of what are now called high-grade epithelial abnormalities (formerly carcinoma-in-situ) could be studied; these women were 25 times more likely than the first group to develop invasive cancer, from which some of them died. Green was accused of callously disregarding the health of these women to satisfy his own scientific curiosity, and condemnation of his actions was central to the findings of the inquiry. However, as Linda Bryder painstakingly demonstrates, there was no such “experiment”. The two groups were a statistical construction in a 1984 paper by four of Green’s colleagues.1 Both groups had been treated but when positive cytology persisted, women were (unsurprisingly) more likely to develop invasive cancer. Green himself certainly advocated a less interventionist approach to the management of cervical pre-malignancy, which in the 1960s tended toward hysterectomy and in the 1970s to cone biopsy. A more low-key approach, now that we understand the role of human papillomavirus and have LLETZ (large loop excision of the transformation zone) procedures available, is universally practised — but, as I’ve observed over my own 30 years of gynaecological practice, it has required a great deal of clinical and scientific effort, much of it by people like Green, to reach this point. Bryder’s book is meticulously referenced and even-handed — although she does not conceal criticism of the feminist health activists who instigated the inquiry. Undoubtedly, the book will arouse much comment from those still able to recall these events, but as a work of medical history it will be a valuable and lasting resource.

Caroline M de Costa

Waiting room. A memoir. Gabrielle Carey. Melbourne: Scribe, 2009 (217 pp). ISBN 9781921372629. I’ve got mixed feelings about memoirs. Having cringed my way through the insipid or vitriolic ramblings of a number of previously admired politicians, actors and journalists, I’ve come to the conclusion that some people’s inner worlds are best kept that way. Nonetheless, I had high hopes as I read the blurb for well known Australian writer Gabrielle Carey’s Waiting room, and plucked it from the review pile to read over a long weekend. All doctors should read patient stories and Carey is an accomplished Australian writer of my generation — surely there would be much in the book to learn from and relate to. I was not disappointed. Waiting room details the diagnosis and management of Carey’s elderly mother’s meningioma. As a medical story, it appears accurate and straightforward and its portrayal of the clinical encounters, surgery and hospital stay is surprisingly undramatic. Without labouring the point, and with a very Australian sense of acceptance and ultimate trust in the medical system’s competence, Carey documents long hours spent waiting for life-changing medical appointments and the all-too-common experience of having planned surgery delayed by a lack of hospital beds. The medical characters are in the book for their functionality rather than their personalities. In fact, the medical story is really a backdrop to the main business of the book; Carey’s relationship with her enigmatic mother. When you reach a certain age, you realise that life is filled with unanswered questions, incomplete understandings, ambiguous meanings and relational loose ends. The beauty of a memoir is that it can unapologetically reflect this chaos without the need for momentous conversations, dramatic revelations and eventual resolution. As a doctor, writer, daughter, wife and mother, I found much to learn from and relate to in Carey’s Waiting room. Sometimes the truth may be stranger than fiction, but generally it is just much more real.

Ruth M Armstrong

Objective: To examine how performances by the Staff Christmas Choir of the Peter MacCallum Cancer Centre (“Peter Mac”) affected inpatients, outpatients and visitors in 2008.Design, setting and participants: During the Christmas season 2008, the Peter Mac Staff Christmas Choir gave seven performances at the Centre. Locations included inpatient wards, outpatient waiting areas and a cafeteria. To assess their response to the choir, oncology inpatients, outpatients and visitors (including early-departing bystanders) were given anonymous, semi-structured questionnaires during and after performances. To analyse the responses, we used a constructivist research approach informed by grounded theory.Main outcome measures: Participants’ descriptions of the choir’s effects on them.Results: Questionnaires were returned by 111 people. The performances were received favourably by 93.7% of respondents, including nine from Jewish, Hindu or atheist backgrounds. Many said the music aroused positive emotions and memories. Several described transformative thoughts and physical reactions, felt affirmed by the Christmas spirit or message, and/or appreciated the peaceful or enlivened and social atmosphere. The choir also elicited personal perspectives about Christmas and Judaism, and the importance of “enjoying the moment”. Only three respondents (2.7%) reported adverse effects, relating to emotional and audible intrusiveness.Conclusions: The Staff Christmas Choir created a supportive and uplifting atmosphere for many oncology patients and their visitors. However, responses from people from non-Christian backgrounds were limited, and further investigation is warranted to extend our understanding of the effect of Christmas music in Australian public health settings.

Clare C O’Callaghan PhD · Colin J Hornby FIR, MAppSc, GradDipEd · Elizabeth J Pearson BAppSc(OT) · David L Ball MB BS, MD, FRANZCR

To the Editor: In 2008, considerable publicity was given to the diagnosis and treatment of prostate cancer in two Australian celebrities: retired Australian Football League player Sam Newman and radio announcer Alan Jones. Newman’s cancer was reported on 5–9 March 2008, culminating in a 12-minute item on the high-rating Australian television program 60 Minutes, when Newman said, “Be tested and enjoy life”. Jones’s announcement of his forthcoming surgery was reported on 3–4 July 2008, and his recovery on 16–27 July. We investigated whether this publicity was associated with changes in the rate of prostate specific antigen (PSA) testing. We used Medicare Australia data on claims for Medicare Benefits Schedule item 66655 (which allows “one annual PSA test for an individual without prior prostate disease”) to determine the mean number of PSA tests and age-standardised testing rate between January 2002 and March 2009. We fitted a linear regression model to the log of the age-standardised rates, modelling secular trends using natural splines with two degrees of freedom (splines were piecewise cubic within the data range and otherwise linear), with indicators set for the average effect of the second, third and fourth quarters and an indicator for the second quarter of 2008. We calculated predicted values and prediction intervals (which take into account both uncertainty in the predicted mean and individual variation), excluding observed data for the second quarter of 2008 (Box). The mean monthly number of PSA tests was 72 064 in the 12 months before April 2008, increasing to 107 633 in April 2008, and 90 813 in July 2008. The seasonally adjusted age-standardised rate of PSA tests rose 17% above the secular trend in the second quarter of 2008 (95% CI, 8%–27%), with the observed data point falling outside the prediction interval (Box). This represented an additional 39 000 PSA tests during the second quarter of 2008. These data show that prostate cancer screening increased significantly in the quarter after media coverage of Newman’s prostate cancer diagnosis. Screening seemed to have returned to the expected seasonal rate by the time of Jones’s diagnosis, in the third quarter of 2008. The impact of celebrity cancer diagnoses on population screening has been demonstrated previously, most recently by an increase in breast cancer screening following news of pop singer Kylie Minogue’s breast cancer diagnosis in 2005.2 In a survey of US men, almost a third stated they would be more likely to have a PSA test after hearing celebrities endorse prostate cancer screening.3 In addition, coverage of melanoma by 60 Minutes in 1987 was followed by a 167% increase in melanoma detection in the subsequent 3 months. Whether the additional 39 000 PSA tests performed in the second quarter of 2008 result in a surge in diagnoses of prostate cancer remains to be seen. Despite continuing uncertainty about the benefits of PSA testing, but unequivocal evidence of harms,,6 Australian men’s willingness to be screened for prostate cancer appears to have been boosted by this celebrity endorsement. Age-standardised* prostate-specific antigen testing rates in Australia (assessed quarterly) Prostate-specific antigen (PSA) testing rates showed a consistent seasonal pattern, with the highest rates in the third quarter (July–September) for all years except 2008, when they were highest in the second quarter. * Standardised to Australian 2001 population.

David P Smith · Mark S Clements · Melanie A Wakefield · Simon Chapman

Currently available, relatively safe vaccines may help tackle this serious and increasing public health problem Melanoma is a serious public health problem in many countries throughout the world, with an incidence increasing at a faster rate than that of any other cancer except lung cancer among women.1 In Europe and the United States, the incidence increased threefold between 1970 and 2000, although improved awareness meant that the mortality rate did not rise so steeply. The risk of melanoma varies greatly (around 100-fold) from region to region, with the highest risk being in Australia where, in 2003, the annual risk was 46.9 per 100 000 population, with estimated lifetime risks of one in 28 and one in 19 up to the ages of 75 and 85 years, respectively.2 With the exclusion of non-melanotic skin cancers, melanoma was the fourth most prevalent cancer in Australia, accounting for 10% of all cases, and the age-standardised risk of melanoma increased by 14% between 1993 and 2003. There are anecdotal reports of regression and even complete resolution of melanoma in patients who had developed serious febrile infections.3 Although these cases are rare and perhaps coincidental, they encouraged us to conduct epidemiological studies on the impact of prior infectious diseases and vaccinations on the risk of melanoma. We therefore established the Febrile Infections and Melanoma (FEBIM) working group in six European countries and Israel, within the Epidemiological Section of the Melanoma Cooperative Group of the European Organization for Research and Treatment of Cancer (EORTC). In an evaluation of our EORTC case–control study, the FEBIM group established that a history of severe but increasingly uncommon infections, with fever above 38.5°C, including sepsis, pneumonia, pulmonary tuberculosis and Staphylococcus aureus infection, was associated with a reduced risk of melanoma.4 Moreover, in patients with histories of severe infections, the extent of risk reduction was directly related to the number of infections. Thus those with histories of one, two to three, and four or more infections had odds ratios of 0.66, 0.63 and 0.32, respectively, and this trend was statistically significant (P = 0.004). It was also established that vaccination early in life against smallpox (with vaccinia vaccine) or tuberculosis (with BCG vaccine), or both, conferred a significant and enduring degree of protection against melanoma. Adjusted odds ratios were 0.40 (95% CI, 0.18–0.85) for BCG alone, 0.60 (95% CI, 0.36–0.99) for vaccinia alone and 0.41 (95% CI, 0.25–0.67) for both vaccines. Not only did these vaccinations afford protection against melanoma, it was subsequently established in an EORTC cohort study of patients with melanoma, that those who developed melanoma had a significantly better prognosis if they had received one or both vaccinations, or had previously had serious but uncommon infections.5 Although confirmatory studies in different settings are required, the multicentre nature of our studies, the high degree of internal consistency of the observations, and the close parallels between their retrospective and prospective arms enhance confidence in the findings. A clue as to how certain vaccinations and infections might protect against melanoma came from the finding that a patient recruited for a trial of a melanoma vaccine had an expanded population of CD8+ T cells that recognised an epitope coded for by a human endogenous retrovirus of the HERV-K family.6 These viruses entered the human germ line millions of years ago and, although no longer capable of replication, can still code for gene products. A possible role of their gene products in the development of melanoma has been the subject of recent research.7 The HERV-K-MEL peptide is expressed on the surface of most human melanomas,6 and it was therefore postulated that a major component of the observed protection by certain infections and vaccinations is the generation of populations of CD8+ T cells cross-reacting with this epitope.8 Accordingly, amino acid sequences with homologies to HERV-K-MEL were sought among pathogens and vaccines, and were found in those associated with protection, but not in those that did not confer protection. A structurally similar sequence was also found in the 17D yellow fever vaccine, suggesting that it might likewise confer protection, and this prediction was confirmed in a recent pilot study of 28 000 adults vaccinated with this vaccine; this pilot study also showed that there was a period of around 10 years between vaccination and observed protection, indicating that the induced immune response is most effective at the time of tumour initiation, which may precede clinical manifestation by several years.9 This protection seems, therefore, to result from prevention of tumour initiation rather than the killing of melanoma cells already present in the body. So, it is important to emphasise that, in contrast to their likely preventive properties, BCG and vaccinia vaccines are of little or no value in treating established melanomas.4,7 Treatment of melanoma by immune modulation is much more complex than prevention, although several approaches of greatly varying complexity hold out hope for effective immunotherapeutic strategies in the future.10 However, our FEBIM studies indicate that certain currently available and relatively well tolerated vaccines against infectious diseases are able to make a significant impact on the serious and increasing public health problem of melanoma. Although it is unlikely that smallpox vaccination will be re-introduced for this purpose, the increasing global incidence of tuberculosis, including extreme drug-resistant forms, makes an additional case for considering widespread neonatal BCG vaccination. Yellow fever vaccine is another possibility as it is cheap and safe, although its efficacy in preventing melanoma requires confirmation in more extensive studies. The currently available evidence indicates the need for further studies to determine whether modifications of vaccination programs to include strategies to reduce the risk of melanoma would be of benefit, especially in high-risk countries such as Australia.

John M Grange MSc, MD · Bernd Krone MD PhD · Klaus F Kölmel MD · Giuseppe Mastrangelo MD

Objective: To examine the initial impact of the National Bowel Cancer Screening Program (NBCSP), which was launched in May 2006 and offers faecal occult blood testing to Australians aged 55 or 65 years.Design and setting: Review of data on colorectal cancer (CRC) cases diagnosed between May 2006 and June 2008 from a prospective database used at 19 Australian hospitals, linked and analysed by BioGrid Australia.Main outcome measures: Number of CRC cases detected through the NBCSP or symptomatic presentation, and differences by sex, stage at diagnosis, tumour location and level of socioeconomic disadvantage.Results: 1628 cases of CRC were identified; 1268 had information on the patients’ test status as part of the NBCSP, and 40 of these (3.2%) were recorded as being detected by the NBCSP. Of 75 CRC cases in patients aged 55 or 65 at diagnosis, 22 were NBCSP-detected. Overall, there was no difference in NBCSP-detected cases by sex. The distribution of tumour locations was similar between NBCSP-detected cases and symptomatic cases, but NBCSP-detected cancers were diagnosed at an earlier stage than symptomatic cancers (stage I, 40% v 14%; stage IV, 3% v 15%, respectively). Of patients diagnosed through the NBCSP, 63% were from areas of least socioeconomic disadvantage (deciles 8–10) and 18% were from the most disadvantaged areas (deciles 1–4) (P = 0.0375).Conclusion: Initiation of the Australian NBCSP has had a measurable impact on CRC stage at diagnosis, and an improvement in survival would be anticipated. The lower uptake among people from disadvantaged areas is of concern.

Sumitra S Ananda MB BS, FRACP · Stephen J McLaughlin MB BS, FRACS, FRCS · Frank Chen MB BS, FRACS · Ian P Hayes MB BS, FRACS, FRCS · Andrew A Hunter MB BS, FRACS · Iain J Skinner MB BS, FRACS · Malcolm C A Steel MB BS, FRACS · Ian T Jones MB BS, FRACS, FRCS · Ian A Hastie MB BS, FRACS · Nicholas A Rieger MB BS, FRACS · Susan Shedda MB BS, FRACS, MPH · Daniel J Compston BHIthInfoManagement · Peter Gibbs MB BS, FRACP, MD

Objective: To investigate trends in the incidence of adenocarcinoma (AC) of the oesophagus in New South Wales, factors associated with a diagnosis of AC, and factors associated with survival of patients with AC.Design and setting: We examined all cases of invasive oesophageal cancer recorded in the NSW Central Cancer Registry from 1972 to 2005. The Accessibility/Remoteness Index of Australia was used to assess geographical remoteness and the Index of Relative Socio-Economic Disadvantage to assess socioeconomic status.Main outcome measures: Incidence of AC; factors associated with diagnosis of AC and survival of patients with AC.Results: The overall incidence of oesophageal AC in NSW increased in both males and females (annual percentage change, 4.2% [95% CL, 2.7%, 5.8%] in males [1988–2005] and 4.3% [95% CL, 1.8%, 7.0%] in females [1983–2005]). A diagnosis of AC was significantly associated with being male (adjusted odds ratio [AOR], 4.37 [95% CL, 3.84, 4.98]; P < 0.001); a younger age at diagnosis (P trend < 0.001); having distant rather than localised disease spread (AOR, 2.12 [95% CL, 1.82, 2.48]; P < 0.001); higher socioeconomic status (P trend < 0.001); and living in an inner regional area (AOR, 1.26 [95% CL, 1.11, 1.43]; P < 0.001) or outer regional area (AOR, 1.19 [95% CL, 1.00, 1.41]; P = 0.05) compared with a major city. Early diagnosis of AC was associated with substantial improvement in survival outcomes: patients with metastatic disease at diagnosis had a three times greater risk of dying than those with localised AC at diagnosis.Conclusion: The incidence of AC is increasing in NSW. Possible contributing factors include increasing obesity, which is associated with increased incidence of gastro-oesophageal reflux disease. Survival may be improved by diagnosis at an earlier stage and changes in modifiable risk factors (eg, smoking, diet, exercise).

Efty P Stavrou PhD · Heather J McElroy MMedStat · Deborah F Baker MPH · Garett Smith BMed, MS, FRACS · James F Bishop MMed, MD, FRACP

Objective: To determine whether a surveillance program including clinical breast examination (CBE) and three screening modalities (magnetic resonance imaging [MRI], high-resolution ultrasound [U/S] and mammographic x-ray [MMX]) was feasible, and whether it could improve detection of pathological lesions in young women at high risk of developing breast cancer.Design, setting and participants: Western Australian women aged 50 years or under at high risk of developing breast cancer were recruited to our study. For a 2-year period, they were offered breast MRI and U/S scans in addition to their annual MMX and CBE. Our study was conducted between June 2002 and October 2005.Main outcome measures: Number and type of cancerous or precancerous lesions; recall rates after screening; comparative sensitivity of screening modalities.Results: Of 102 women approached, 72 agreed to participate. Fifteen lesions were detected, of which three were significant: a metastatic papillary cancer in an axillary lymph node, a borderline lesion (multiple papillomatosis with atypia), and a papilloma. All 15 lesions were visible on MRI, and four were detected by MRI only. Only one lesion was visible on all three imaging modalities. Nothing significant was detected by CBE. The recall rate after MRI scans fell from 9/72 (12.5%) in the first year to 5/67 (7.5%) in the second year.Conclusion: Our study gave valuable experience in a team approach to screening MRI, and showed that MRI can detect more lesions than MMX or U/S in women at high risk of developing breast cancer. Screening U/S may not add value to MMX and MRI screening, and we suggest a single oblique-view MMX may be used in some cases.

Christobel M Saunders MB BS, FRCS, FRACS · Gudrun Peters FRANZCR · Glenys Longman RN · Jacqueline Thomson MB ChB, FRANZCR · Donna Taylor MB BS, FRANZCR · Jianmin Hua BSc · Michelle Bennett MB ChB, FRANZCR · Elizabeth Wylie MB BS, FRANZCR · Jack Goldblatt MB BS, MD, FRACP · Arlene Chan MB BS, FRACP, MMed · James Anderson MB BS, FRANZCR

Objective: To evaluate the difference in time taken for ethics and site governance approval for multicentre clinical trials using two different systems of ethics review.Design: We evaluated the times to final ethics and governance approval for two international, multicentre clinical trials of treatment for metastatic colorectal cancer: the MAX trial, using a non-centralised ethics review system, and the CO.20 trial, using the new New South Wales centralised ethics review system.Main outcome measure: Time from trial submission to overall study approval.Results: The median time taken to obtain ethics approval for the MAX trial at 16 NSW sites was 100 days (range, 36–161 days). The median time to obtain central ethics approval for the CO.20 trial at 14 NSW sites was 77 days, with an additional 60 days (range 20–79 days) required to obtain site-specific research governance approval.Conclusions: Any difference in time to approval between the review systems was outweighed by the overall time taken. However, the time spent by both the coordinating centre and local sites in collation, submission and correspondence was greatly reduced, and the centralised process allowed for standardised documentation at all study sites.

Sian C Hicks BSc(Hons), PhD · Rebecca E James BA/BSc, GradDipEd, MScMed · Nicole Wong RN, BN, BSc(Hons) · Niall C Tebbutt BM BCh, PhD, FRACP · Kate Wilson BA, MPH

To the Editor: We applaud the attempt by Boyes and colleagues to ascertain the level of psychological distress experienced by patients over the years following diagnosis with cancer, through a retrospective, cross-sectional survey of New South Wales cancer registrants.1 However, we believe several methodological limitations ought to reduce the confidence with which the authors drew their conclusions. The authors’ comment that “life after cancer is not all doom and gloom” was, perhaps, intended to be a little facetious. From a scientific point of view, however, such a statement is also very difficult to ever disprove — of course it isn’t all doom and gloom. Further, the authors’ assertion that psychosocial wellbeing several years after cancer diagnosis is comparable with that of the general population cannot be substantiated by studies conducted by this method. As the survey was cross-sectional, we have no information about the level of distress experienced during the years since diagnosis. A longitudinal design is now de rigueur in this field for this reason. Certainly, clinical experience shows us that many patients actually do experience “doom and gloom” often, indeed, arising from “insidious and relentless disease” — this group may not, however, be well represented by a cancer registry survey sampling 5 years after diagnosis. Selection bias is a major problem. Of the sample of 2029 eligible, randomly selected people, 655 (32%) were deemed ineligible, with one of the exclusion criteria being not “mentally capable of participating”. Could this sizable subgroup have included those who were distressed? Also, of the eligible sample of 1374, 366 (27%) declined to participate. What were the reasons for refusal? Were some too distressed to participate? In summary, how were the 37% who did not participate in the study faring 5 years after diagnosis? The Hospital Anxiety and Depression Scale (HADS), used by Boyes et al to detect psychological distress in early-stage breast cancer, under-reports distress when recommended cut-off scores are used, compared with a structured clinical interview validated to provide Diagnostic and statistical manual of mental disorders, fourth edition (DSM-IV2) diagnoses.3 In other words, the HADS is known to lack sensitivity and positive predictive power in the cancer setting. Other cancer researchers found the same.4,5 Further, measuring distress only in terms of anxiety and depression 5 years into the adjustment process fails to capture the quality of continuing distress and the degree to which traumatic growth and other forms of meaning-based adjustment have been achieved. The limitations of the HADS should have been better acknowledged.

Jeremy W Couper · Anthony W Love · Annabel C Pollard · Sidney Bloch

In reply: We thank Couper and colleagues for their interest in our article,1 but reiterate that we focused on anxiety and depression experienced by long-term cancer survivors, specifically at 5–6 years after diagnosis. We agree that longitudinal studies are vital for understanding the level of psychological distress experienced during the years since diagnosis, and as discussed in our article, we are currently undertaking a longitudinal study with a diversity of cancer patients to assess a comprehensive range of physical, psychological, social and lifestyle effects of cancer. Although opinions on the performance of the Hospital Anxiety and Depression Scale (HADS) vary, it is one of the most popular measures of psychological distress, and has been used extensively across the cancer continuum.2 A recent review of the validity of the HADS concluded that it performs well in screening for caseness of anxiety disorders and depression in a range of patient populations, including patients with cancer, and in the general population.3 We acknowledge that our study had the strengths and limitations normally associated with recruiting through a population-based cancer registry.4 Nevertheless, our results are consistent with a growing body of evidence indicating that most cancer survivors are doing well 5 or more years after diagnosis.5 We believe our conclusions are justified.

Allison W Boyes · Afaf Girgis · Alison C Zucca · Christophe Lecathelinais

A 65-year-old woman underwent integrated fluorodeoxyglucose positron emission tomography and computed tomography (FDG-PET/CT [Philips GXL, Philips Medical Systems, Milpitas, Calif, USA]) to stage a newly diagnosed squamous cell carcinoma of the tongue (Figure, A). The scan revealed two additional, unexpected synchronous tumours, one in the left axilla (B) and the other in the sigmoid colon (C). The patient underwent subtotal glossectomy, after which further investigations confirmed a node-positive neuroendocrine carcinoma of the left breast and a dysplastic colonic tubulovillous adenoma. The detection of three synchronous tumours of different aetiology in the one patient on PET is rare. Previous cases of synchronous tumours detected on PET involved tumours of the head and neck, upper gastrointestinal tract and lungs, thought to be related to shared risk factors, such as smoking.1

Paul M Leong · Michael Lin · Allan R Fowler

Large-scale randomised controlled trials in Europe and the United States are informing evidence-based clinical advice Population-wide prostate-specific antigen (PSA) screening for prostate cancer is not recommended in Australia, primarily because of the lack of large-scale randomised trial evidence of a beneficial effect on prostate cancer mortality and the known harms of overdiagnosis and unnecessary treatment that may ensue. In spite of this, PSA testing is common; the limited evidence available suggests that more than 50% of Australian men over the age of 50 years have had the test.1,2 In March 2009, the New England Journal of Medicine published the first articles from two large, randomised controlled trials to report on whether screening for PSA reduces the risk of death from prostate cancer.3,4 It was hoped that these long expected — albeit not final — reports from the European Randomized Study of Screening for Prostate Cancer and the prostate cancer component of the US Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (Box) would answer outstanding questions about the efficacy of prostate cancer screening. In the European trial, 162 243 men aged 55–69 years were randomly allocated to receive an invitation to undergo PSA screening about once every 4 years or usual care. After a median of 9 years of follow-up, there were 71% more cases of prostate cancer and a 20% reduction in the risk of death from prostate cancer in those allocated to screening than in those receiving usual care (relative risk [RR], 0.80; 95% CI, 0.65–0.98), but no reduction in overall mortality (RR, 0.99; 95% CI, 0.97–1.02). In the US trial, 76 693 men aged 55–74 years were randomly allocated to receive an annual PSA test for 6 years and an annual digital rectal examination for 4 years, or usual care. After a minimum of 7 years of follow-up, there were 22% more cases of prostate cancer but no reduction in the risk of death from prostate cancer in those randomly allocated to screening compared with usual care (RR, 1.13; 95% CI, 0.75–1.70). The trials have a number of notable differences, including length of follow-up, screening interval, PSA threshold for recommending biopsy, and proportion of the control group screened (Box). Although these differences prevent easy comparison of their results, it is important to note that the relative risks for prostate cancer mortality with screening do not differ significantly between the trials (χ12 = 2.19, P = 0.14), and that the 95% confidence interval about their summary relative risk, estimated using fixed-effects meta-analysis, includes unity (summary RR, 0.86; 95% CI, 0.71–1.03) (Box). These trials have important limitations. First, their combined power is low, so that a moderate beneficial effect of screening cannot be established or excluded without longer follow-up or other trials. Moreover, in the US trial, although 85% of the screening arm had a PSA test, so too did 52% of the control arm. With only a 33% difference in screening between the groups, the power of the study to detect an effect of screening was substantially reduced. Screening in the control arm of the European trial was not reported. Second, the trials are yet to report on other outcomes of PSA screening, including major treatment side effects, which are crucial to deciding whether large-scale screening is appropriate. The risk–benefit equation for PSA screening is complex. The major potential benefit is a reduction in risk of death from prostate cancer. The main risks are physical or psychological complications of screening and detection, overdiagnosis and concomitant treatment side effects, and unwarranted costs. Some elements of these risks were reported in the European trial. Overall, 16.2% of all PSA tests gave positive results and presumably led to biopsy, and 75.9% of these were found to be false-positives.4 The risk of undergoing radical prostatectomy in the screening group was nearly three times that in the control group (277 versus 100 per 10 000 men).5 To prevent one prostate cancer death, 1410 men needed to be screened, and 48 men needed to be treated.4 Although the number needed to screen compares favourably with that for breast6 and colorectal7 cancer, the issues around overdiagnosis are probably greater for prostate cancer. The cost-effectiveness of PSA screening is yet to be established. Even if it were shown to be favourable, the appropriate testing interval, target age range and PSA cut-off levels are not known. Hence, the persisting uncertainty about the effect of PSA testing on mortality is compounded by other outstanding questions. What are the implications of the trial results for men and their doctors? The findings of these studies are finely balanced. Before they were reported, we had no valid randomised trial evidence; now we have evidence, but it is inconclusive and consistent with a modest but uncertain reduction in prostate cancer mortality at 10 years. Both trials can continue to inform us with longer-term follow-up and comprehensive assessment of all treatment outcomes. What is the best advice that we can give now about PSA screening for prostate cancer? Population-wide PSA screening should not be recommended. The findings from these two studies provide important evidence but are not adequate to allow a conclusion either for or against a benefit of screening. In addition, they reveal potentially high levels of overdiagnosis and its consequent costs and harm. For individual men considering PSA screening, the potential risks and benefits should be carefully considered. Information from these trials and other studies, and data on the potential of treatments for localised prostate cancer to affect quality of life should be communicated, stating that uncertainty continues to cloud this issue. Clearly, men with a life expectancy of less than 10 years are particularly unlikely to benefit from screening, but those who wish to be tested after considering the evidence available should be given the test. Clinical practice includes treating patient concerns and worries, so that if a well informed man really wants to know whether he has prostate cancer, proceeding with PSA testing is a reasonable step. GPs play a crucial role in disclosing the uncertainties of PSA testing and in supporting men in their decision making. Research assumes even greater importance when decision making is clouded by uncertainty. There is a clear need for more evidence on all effects of PSA testing. A greater understanding of decision making around PSA testing is also needed,8 as is better information to help men and their doctors address the issue. Whether screening for prostate cancer lowers risk of death from prostate cancer remains uncertain. However, our uncertainty is increasingly based on evidence, which is far preferable to uncertainty based on ignorance. Characteristics of two published randomised controlled trials of the effect of PSA screening on prostate cancer mortality ERSPC (n = 162 243) PLCO (n = 76 693) Trial period 1991–2003 1993–2001 Age range (years) 55–69 55–74 No. of participants (S v C) 72 890 v 89 353 38 343 v 38 350 Average screening interval 4 years 1 year PSA cut-off for biopsy 3 ng/mL* 4 ng/mL Tested with PSA (S v C) 82% v not stated 85% v 52% Prostate cancers (no.) 5990 v 4307 2820 v 2322 Cumulative incidence of prostate cancer (S v C) 8.2% v 4.8% 7.4% v 6.1% RR (95% CI) of incident prostate cancer in S v C 1.71 (1.64–1.78)† 1.22 (1.16-1.29) No. of prostate cancer deaths (S v C) 214 v 326 50 v 44 Prostate cancer death rate (per 10 000 person-years) (S v C) 3.3 v 4.1 2.0 v 1.7 RR of prostate cancer death in S v C (95% CI) 0.80 (0.65–0.98) 1.13 (0.75–1.70) Summary RR of prostate cancer death in S v C (95% CI) 0.86 (0.71–1.03)‡ PSA = prostate-specific antigen. ERSPC = European Randomized Study of Screening for Prostate Cancer. PLCO = prostate cancer component of the US Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial. S = screening group. C = control group. RR = relative risk. * Two centres had slightly different cut-offs, of around 4 ng/mL. † Calculated from data presented in the report of the ERSPC, Table 1, using a Poisson approximation for the variance. ‡ χ2 (heterogeneity) = 2.19, P = 0.14.

David P Smith BA, MPH, PhD · Emily Banks MB BS(Hons), PhD, FAFPHM · Mark S Clements BSc(Hons), PhD · Robert A Gardiner MD, FRCS, FRACS · Bruce K Armstrong MB BS(Hons), DPhil, FAFPHM

Objective: To investigate the frequency, goals and outcomes of second-opinion consultations at the Sydney Cancer Centre.Design, setting and participants: A questionnaire-based study of patients who registered to see a medical oncologist at the Sydney Cancer Centre between January 2006 and January 2008 and who were seeking a second opinion.Main outcome measures: Proportion and demographic characteristics of patients who had previously seen a medical oncologist and who stated they were seeking a second opinion.Results: 123 of 1892 new patients (6.5%) stated that they were seeking a second opinion, of whom 22 declined study participation, were excluded from study participation or had been referred specifically for enrolment in a particular clinical trial. Of the remaining 101 patients, 77 completed a questionnaire; 59 were women and 26 had a university degree. Reasons for seeking second opinions included: to obtain information related to treatment (54 patients), for reassurance about diagnosis or treatment (47), and dissatisfaction with the information given by the first medical oncologist (24). Sixty-four patients reported that they received new information at the second-opinion consultation, with 45 identifying discussion of treatment options and 34 identifying discussion of future or prognosis. Fifty-one patients reported how the second-opinion consultation differed from the first, identifying it as longer (24), and indicating that the oncologist answered concerns (26). Most patients were aware of multidisciplinary teams and treatment guidelines, but fewer had read guidelines.Conclusions: Patients seeking a second opinion from a medical oncologist are typically more educated, younger and female, probably due to preferences for more detailed information. The most common reasons for seeking a second opinion were to obtain additional information or reassurance about recommended management.

Martin H N Tattersall MD, MSc, FRACP · Rachel F Dear MB BS, FRACP · Jesse Jansen MA, PhD · Heather L Shepherd PhD · Rhonda J Devine RN · Lisa G Horvath MB, FRACP, PhD · Michael J Boyer MB, FRACP, PhD

The best available evidence for making decisions about prostate-specific antigen (PSA) screening comes from two recent randomised trials, the larger and more robust of which showed that PSA screening reduced the risk of death from prostate cancer, but that the absolute benefit was small, and the chance of prostate cancer being diagnosed and treated (even if biologically unimportant) was increased by a much larger amount. The important question is whether the small reduction in numbers of deaths outweighs the harms inherent in the diagnosis and treatment of many additional cancers. Men considering screening should understand both its possible benefit and its possible harms, and that the harms are more immediate than any benefit. The challenge for future research is to find a test that reliably detects prostate cancers that are curable if they are treated early and life-threatening if they are not.

Alexandra L Barratt MB BS, MPH, PhD · Martin R Stockler MB BS, MSc(ClinEpi), FRACP