A study of margin width and local recurrence in breast conserving therapy for invasive breast cancer.

PURPOSE: Debate continues on what is an adequate margin width to define a clear margin and whether there is a need to excise pectoral fascia or remove skin in breast conserving surgery. This study set out to provide answers to these questions. PATIENTS AND METHODS: 1411 patients with invasive breast cancer were treated by breast conserving surgery and post-operative whole breast radiotherapy from January 2000 to December 2005. Distance from each margin to any in situ or invasive cancer was measured and recorded. If full thickness of breast tissue was removed no re excision of anterior and posterior margins was performed even if disease was <1 mm from a margin. Patients ≤50 years of age and those with anterior or posterior margins <1 mm to invasive cancer had a radiation boost. Median follow-up time was 6.4 years. RESULTS: Local in breast tumour relapse (IBTR) occurred in 50 patients. The overall actuarial IBTR rate at 5 years was 2.2%. There was no difference in IBTR when comparing patients with radial margins of 1-5 mm or 5-10 mm. Anterior and posterior margins <1 mm or with ink on tumour cells were not associated with an increase in IBTR. CONCLUSION: There is no justification for radial margins of greater than 1 mm. If the anterior or posterior margin is <1 mm and full thickness of breast tissue has been removed, then re excision of these margins is unnecessary if boost radiotherapy is delivered.

Pooled long-term outcomes from two randomized trials of axillary node sampling with axillary radiotherapy versus axillary node clearance in patients with operable node-positive breast cancer.

BACKGROUND: The aim was to determine long-term overall, breast cancer-specific and metastasis-free survival as well as axillary relapse rate from a pooled analysis of two randomized trials in women with operable breast cancer. These trials compared axillary node sampling (ANS), combined with axillary radiotherapy (AXRT) if the sampled nodes were involved, with axillary node clearance (ANC). METHODS: Data from two clinical trials at the Edinburgh Breast Unit that randomized patients between 1980 and 1995 were pooled. Long-term survival was analysed using Kaplan-Meier curves and Cox regression, with separate analyses for patients with node-positive (ANS + AXRT versus ANC) and node-negative (ANS versus ANC) disease. RESULTS: Of 855 women randomized, 799 were included in the present analysis after a median follow-up of 19·4 years. Some 301 patients (37·7 per cent) had node-positive disease. There was no evidence of a breast cancer survival advantage for ANS versus ANC in patients with node-negative disease (hazard ratio (HR) 0·88, 95 per cent c.i. 0·58 to 1·34; P = 0·557), or for ANS + AXRT versus ANC in those with node-positive breast cancer (HR 1·07, 0·77 to 1·50; P = 0·688). There was no metastasis-free survival advantage for ANS versus ANC in patients with node-negative tumours (HR 1·03, 0·70 to 1·51; P = 0·877), or ANS + AXRT versus ANC in those with node-positive disease (HR 1·03, 0·75 to 1·43; P = 0·847). Node-negative patients who underwent ANS had a higher risk of axillary recurrence than those who had ANC (HR 3·53, 1·29 to 9·63; P = 0·014). Similarly, among women with node-positive tumours, the risk of axillary recurrence was greater after ANS + AXRT than ANC (HR 2·64, 1·00 to 6·95; P = 0·049). CONCLUSION: Despite a higher rate of axillary recurrence with ANS combined with radiotherapy to the axilla, ANC did not improve overall, breast cancer-specific or metastasis-free survival. Axillary recurrence is thus not a satisfactory endpoint when comparing axillary treatments.

Time dependence of biomarkers: non-proportional effects of immunohistochemical panels predicting relapse risk in early breast cancer.

BACKGROUND: We investigated the impact of follow-up duration to determine whether two immunohistochemical prognostic panels, IHC4 and Mammostrat, provide information on the risk of early or late distant recurrence using the Edinburgh Breast Conservation Series and the Tamoxifen vs Exemestane Adjuvant Multinational (TEAM) trial. METHODS: The multivariable fractional polynomial time (MFPT) algorithm was used to determine which variables had possible non-proportional effects. The performance of the scores was assessed at various lengths of follow-up and Cox regression modelling was performed over the intervals of 0-5 years and >5 years. RESULTS: We observed a strong time dependence of both the IHC4 and Mammostrat scores, with their effects decreasing over time. In the first 5 years of follow-up only, the addition of both scores to clinical factors provided statistically significant information (P<0.05), with increases in R(2) between 5 and 6% and increases in D-statistic between 0.16 and 0.21. CONCLUSIONS: Our analyses confirm that the IHC4 and Mammostrat scores are strong prognostic factors for time to distant recurrence but this is restricted to the first 5 years after diagnosis. This provides evidence for their combined use to predict early recurrence events in order to select those patients who may/will benefit from adjuvant chemotherapy.

The clinical effectiveness and cost-effectiveness of different surveillance mammography regimens after the treatment for primary breast cancer: systematic reviews registry database analyses and economic evaluation.

BACKGROUND: Following primary breast cancer treatment, the early detection of ipsilateral breast tumour recurrence (IBTR) or ipsilateral secondary cancer in the treated breast and detection of new primary cancers in the contralateral breast is beneficial for survival. Surveillance mammography is used to detect these cancers, but the optimal frequency of surveillance and the length of follow-up are unclear. OBJECTIVES: To identify feasible management strategies for surveillance and follow-up of women after treatment for primary breast cancer in a UK setting, and to determine the effectiveness and cost-effectiveness of differing regimens. METHODS: A survey of UK breast surgeons and radiologists to identify current surveillance mammography regimens and inform feasible alternatives; two discrete systematic reviews of evidence published from 1990 to mid 2009 to determine (i) the clinical effectiveness and cost-effectiveness of differing surveillance mammography regimens for patient health outcomes and (ii) the test performance of surveillance mammography in the detection of IBTR and metachronous contralateral breast cancer (MCBC); statistical analysis of individual patient data (West Midlands Cancer Intelligence Unit Breast Cancer Registry and Edinburgh data sets); and economic modelling using the systematic reviews results, existing data sets, and focused searches for specific data analysis to determine the effectiveness and cost-utility of differing surveillance regimens. RESULTS: The majority of survey respondents initiate surveillance mammography 12 months after breast-conserving surgery (BCS) (87%) or mastectomy (79%). Annual surveillance mammography was most commonly reported for women after BCS or after mastectomy (72% and 53%, respectively). Most (74%) discharge women from surveillance mammography, most frequently 10 years after surgery. The majority (82%) discharge from clinical follow-up, most frequently at 5 years. Combining initiation, frequency and duration of surveillance mammography resulted in 54 differing surveillance regimens for women after BCS and 56 for women following mastectomy. The eight studies included in the clinical effectiveness systematic review suggest surveillance mammography offers a survival benefit compared with a surveillance regimen that does not include surveillance mammography. Nine studies were included in the test performance systematic review. For routine IBTR detection, surveillance mammography sensitivity ranged from 64% to 67% and specificity ranged from 85% to 97%. For magnetic resonance imaging (MRI), sensitivity ranged from 86% to 100% and specificity was 93%. For non-routine IBTR detection, sensitivity and specificity for surveillance mammography ranged from 50% to 83% and from 57% to 75%, respectively, and for MRI from 93% to 100% and from 88% to 96%, respectively. For routine MCBC detection, one study reported sensitivity of 67% and specificity of 50% for both surveillance mammography and MRI, although this was a highly select population. Data set analysis showed that IBTR has an adverse effect on survival. Furthermore, women experiencing a second tumour measuring >20 mm in diameter were at a significantly greater risk of death than those with no recurrence or those whose tumour was <10 mm in diameter. In the base-case analysis, the strategy with the highest net benefit, and most likely to be considered cost-effective, was surveillance mammography alone, provided every 12 months at a societal willingness to pay for a quality-adjusted life-year of either £20,000 or £30,000. The incremental cost-effectiveness ratio for surveillance mammography alone every 12 months compared with no surveillance was £4727. LIMITATIONS: Few studies met the review inclusion criteria and none of the studies was a randomised controlled trial. The limited and variable nature of the data available precluded any quantitative analysis. There was no useable evidence contained in the Breast Cancer Registry database to assess the effectiveness of surveillance mammography directly. The results of the economic model should be considered exploratory and interpreted with caution given the paucity of data available to inform the economic model. CONCLUSIONS: Surveillance is likely to improve survival and patients should gain maximum benefit through optimal use of resources, with those women with a greater likelihood of developing IBTR or MCBC being offered more comprehensive and more frequent surveillance. Further evidence is required to make a robust and informed judgement on the effectiveness of surveillance mammography and follow-up. The utility of national data sets could be improved and there is a need for high-quality, direct head-to-head studies comparing the diagnostic accuracy of tests used in the surveillance population. FUNDING: The National Institute for Health Research Health Technology Assessment programme.

Trends in classification, referral and treatment and the effect on outcome of patients with glioma: a 20 year cohort.

This retrospective audit was conducted to examine the changes in patient characteristics, referral, treatment and outcome over a 20-year period in a large regional neuro-oncology centre, focusing on the impact of the changes in pathological classification of gliomas. Using the Edinburgh Cancer Centre (ECC) database all cases of glioma were identified and patient, tumour and treatment characteristics noted. Survival was calculated from date of surgery or, if no operation was performed, the date of referral. Comparison was made between four periods 1988-1992 (c1), 1993-1997(c2), 1998-2002(c3) and 2003-2007 (c4). During the 20 years, 1175 patients with a glioma were referred to ECC. The median age increased from 53 years to 57 years (p < 0.001) but the proportion without pathology remained unchanged (10%). The distribution of pathological grades changed over time Grade I-II: 24, 6, 6, and 6%, Grade III: 42, 27, 17, and 13% and Grade IV: 24, 61, 68, and 68% in c1, c2, c3 and c4, respectively (p < 0.001). Immediate RT was given to 68% (c1), 70% (c2), 78% (c3) and 79% (c4). Median interval from resection to RT reduced from 43 days (c1) to 36 days (c4) (p < 0.001). 5-year overall survival for patients with Grade III lesions increased: 21% (c1), 35% (c2), 37% (c3), 33% (c4) as did 1-year overall survival for Grade IV lesions: 18% (c1), 26% (c2), 29% (c3), 27% (c4)). This improvement probably reflects the change in pathological classification rather than a change in management. Proportional hazards analysis of grade IV 1993-2007 only (to reduce pathological variation) showed that younger age, frontal lesions, excision, higher RT dose had reduced hazard of death. Interval from surgery to RT had no impact on survival in this series.

Breast radiotherapy for occult breast cancer with axillary nodal metastases--does it reduce the local recurrence rate and increase overall survival?

AIMS: The optimal management of axillary lymph node metastases from occult breast cancer (TXN1-2M0) is uncertain and practice varies in the use of primary breast radiotherapy. We conducted a retrospective review to examine clinical outcomes for patients managed with or without primary breast radiotherapy. MATERIALS AND METHODS: Case records from the clinical oncology database were reviewed to identify patients presenting with axillary nodal metastases but no detectable primary tumour between 1974 and 2003. Fifty-three patients with TXN1-2M0 breast cancer were identified, representing 0.4% of patients managed for breast cancer during this period. Of those tested, 59% had oestrogen receptor-positive tumours. Seventy-seven per cent received ipsilateral breast radiotherapy. RESULTS: There was a trend towards reduced ipsilateral breast tumour recurrence in patients who received radiotherapy (16% at 5 years, 23% at 10 years) compared with those who did not (36% at 5 years, 52% at 10 years). Similarly, the locoregional recurrence rate was 28% at 5 years for patients who received radiotherapy compared with 53.7% at 5 years for non-irradiated patients. Breast cancer-specific survival was higher (P=0.0073; Log-rank test) in patients who received ipsilateral breast radiotherapy (72% at 5 years, 66% at 10 years) compared with those who did not (58% at 5 years, 15% at 10 years). CONCLUSION: Primary breast radiotherapy may reduce ipsilateral breast tumour recurrence and may increase survival in patients presenting with axillary lymph node metastases and occult breast primary (TXN1-2M0). Larger studies or prospective registration studies are needed to validate these findings.

Three years of erlotinib in routine practice for non-small cell lung cancer in South East Scotland.

We present a review of 111 patients who were treated over an initial 3-year period with erlotinib. The median treatment time was 68 days and 59% of patients had stopped treatment within the first 3 months. However, 20 patients were on erlotinib for more than 12 months. Performance status and smoking history were the significant prognostic factors. The overall 3-year survival in patients who had never smoked was 26%.

Health care costs for the treatment of breast cancer recurrent events: estimates from a UK-based patient-level analysis.

Cost pressures and the need to demonstrate cost-effectiveness of new interventions require consideration of the costs of treating disease. This study presents analyses of resource use data covering 199 postmenopausal women who experienced a breast cancer recurrent event between 1991 and 2004 and were treated at the Western General Hospital, Edinburgh. Aggregate (5-year) treatment costs for alternative recurrent events were estimated, as well as the annual costs incurred by patients experiencing contralateral, locoregional, or distant recurrence, who remained alive without further recurrence for a year. The 95% confidence intervals for the 5-year costs of recurrence ranged from pounds 10,000 to pounds 37,000 for locoregional recurrence, and pounds 14,500- pounds 20,000 for distant recurrence. No evidence of significant variations in these costs across time periods between 1991 and 2004 was identified. Annual costs for patients remaining in the same health state showed high initial costs for contralateral and locoregional recurrence, with low costs in subsequent years, while costs associated with distant recurrence declined at a slower rate and plateaued at 4-5 years post-diagnosis. The cost estimates presented in this paper not only inform the magnitude of the resource consequences of breast cancer recurrences, but they are also better suited to informing cost-effectiveness analyses, which have a far greater role in allocating health-care resources.

Changing pattern of the detection of locoregional relapse in breast cancer: the Edinburgh experience.

The guidelines for follow-up of breast cancer patients concentrate on the first 3-5 years, with either reduced frequency of visits or discharge after this. They also recommend mammography, but no evidence exists to inform frequency. We analyse treatable relapses in our unit from 1312 patients with early stage breast cancer treated by breast conserving surgery (BCS) and postoperative radiotherapy between 1991 and 1998 to assess appropriateness of the guidelines. A total of 110 treatable relapses were analysed. Treatable relapse developed at 1-1.5% per year throughout follow-up. Forty-eight relapses were in ipsilateral breast, 25 ipsilateral axilla, 35 contralateral breast, 2 both breasts simultaneously. Thirty-seven relapses (33.5%) were symptomatic, 56 (51%) mammographically detected, 15 (13.5%) clinically detected, 2 (2%) diagnosed incidentally. Mammography detected 5.37 relapses per 1000 mammograms. Patients with symptomatic or mammographically detected ipsilateral breast relapse had significantly longer survival from original diagnosis (P=0.0002) and from recurrence (P=0.0014) compared with clinically detected. Treatable relapse occurs at a constant rate for at least 10 years. Clinical examination detects a minority (13.5%). Relapse diagnosed clinically is associated with poorer outcome. Long-term follow-up based on regular mammography is warranted for all patients treated by BCS.

Improving outcomes for limited stage small cell lung cancer patients in Scotland with concomitant chemoradiation.

AIMS: To determine whether the introduction of early concomitant chemoradiotherapy for patients with limited stage small cell lung cancer (LS-SCLC) has resulted in acceptable outcomes and toxicity in a UK practice. MATERIALS AND METHODS: The case records of all patients with LS-SCLC treated with chemoradiotherapy from July 2001 to 2004 were reviewed, and subjected to descriptive statistics and proportional hazards analysis. RESULTS: Concomitant chemoradiotherapy was delivered to 30 patients and sequential chemoradiotherapy was delivered to 36 patients. The former patients tended to be younger (mean 58.9 vs 64.1 years, P=0.01); the latter patients tended to have bulkier disease. There was no difference in performance status, but cisplatin was given more often in the former group (90% vs 44%, P<0.0001). Grade 3 acute oesophagitis occurred in less than 10% of either group and there were no cases of grade 3 or greater pneumonitis. Two-year actuarial survival for the concomitant group was 53% (95% confidence interval 36-71%) and 36% (95% confidence interval 20-52%) for the sequential group (P=0.018). Proportional hazards analysis showed an increased hazard of death with increasing performance status and age, sequential therapy and the use of cisplatin with sequential therapy. CONCLUSION: Concomitant chemoradiotherapy can be safely given in a UK population with outcomes comparable with those reported in North American series.

Is routine chest radiography a useful test in the follow up of all adult patients with soft tissue sarcoma?

Following treatment for localized soft tissue sarcoma the risk of relapse is either locally or in the lungs. In Edinburgh patients are reviewed every 6 months with a chest X-ray (CXR). The radiation exposure over a 10 year follow up remains small, but it is unclear if all patients, irrespective of the initial grade of their primary tumour, require this. To determine the pick up rate of lung metastases by routine CXR over a 10 year period and to review the primary histology. Adult patients on routine follow up between 1994 and 2004 were identified and the notes of those with lung metastases reviewed. Data was collected on their initial histology, and date and method of diagnosis of lung metastases. 179 patients were under follow up. 24 (13%) developed lung metastases. For 2, notes were not found. 6 (27%) had metastases diagnosed by routine CXR, 9 (41%) had metastases diagnosed by non routine CXR and 7 (32%) had metastases diagnosed by CT. On review of histology none were grade 1, 4 (18%) were grade 2 and 18 (82%) were grade 3. 155 patients received. 6 monthly CXR for 10 years with no detection of lung metastases. Lung metastases occurred in a minority of patients (13%) and most (82%) occurred in patients with grade 3 tumours. No patients with grade 1 tumours developed lung metastases. Thus routine CXR may be appropriate on grade 3 tumours, but not on lower grade tumours where other risk factors are absent.

The management and outcome of patients with germ-cell tumours treated in the Edinburgh Cancer Centre between 1988 and 2002.

AIMS: The aim of this retrospective analysis was to review the outcome of patients with germ-cell tumours treated in the Edinburgh Cancer Centre over the past 15 years, and to see whether there had been any changes over three 5-year cohorts. MATERIALS AND METHODS: Patients referred with gonadal and extra-gonadal primary germ-cell tumours, between 1988 and 2002, were identified from the departmental database, and survival by stage and prognostic group was analysed. RESULTS AND CONCLUSIONS: The proportion of patients with stage I seminoma has significantly increased. The good prognosis of patients with early stage disease is confirmed, with the outcome for some groups of patients being better than expected. There is a non-significant trend to improved results over the three 5-year cohorts. The outcome for patients with stage IV seminoma is worse than would be expected, but numbers are small. The poor prognosis of patients with non-seminomatous germ-cell tumours who fall into the International Germ Cell Consensus Classification (IGCCC) poor-prognostic group is confirmed. Failure of patients with metastatic non-seminomatous germ-cell tumours to achieve a complete response to initial therapy is shown to be a poor prognostic indicator.

Testicular sex cord-stromal tumours: the Edinburgh experience 1988-2002, and a review of the literature.

AIMS: Sex cord-stromal tumours of the testis are uncommon tumours, accounting for around 5% of testicular neoplasms. Treatment is primarily surgical, with no adjuvant therapy of proven benefit. We present a single-centre experience over a period of 15 years. MATERIALS AND METHODS: From 1988 to 2002, 18 patients with a diagnosis of sex cord-stromal tumour were referred to our centre. A retrospective analysis of their case notes was made and a pathological review undertaken. RESULTS: Sixteen were Leydig-cell tumours and two were Sertoli cell. For the Leydig-cell tumours, the median age at presentation was 42 years, 50% presented with a testicular mass and 31% with gynaecomastia. Two patients followed a malignant course: one revealing disease dissemination at initial staging, and a second 12 months after potentially curative orchidectomy. Salvage retroperitoneal lymphadenectomy in the latter patient proved unsuccessful. Clinical outcome correlated strongly with the presence of adverse pathological features described previously in the literature. After a median follow-up of 46 months, two patients have developed progressive disease, and two patients have died, one of metastatic Leydig-cell tumour. No patient defined as being of low malignant potential on pathological examination has relapsed outside our review period of 2 years. CONCLUSION: We confirm the overall excellent prognosis for most of the patients with sex cord-stromal tumours of the testis. Compared with most previous reports, pathological features seem to predict with reasonable accuracy the risk of malignant behaviour, and can adequately inform the subsequent review policy.

Symptom control and quality of life in people with lung cancer: a randomised trial of two palliative radiotherapy fractionation schedules.

AIMS: To determine whether palliation of chest symptoms from a 10 Gy single fraction (regimen 1) was equivalent to that from 30 Gy in 10 fractions (regimen 2). MATERIALS AND METHODS: Patients with cytologically proven, symptomatic lung cancer not amenable to curative therapy, with performance status 0-3, were randomised to receive either 30 Gy in 10 fractions or a 10 Gy single fraction. Local symptoms were scored on a physician-assessed, five-point categorical scale and summed to produce a total symptom score (TSS). This, performance status, Hospital Anxiety and Depression (HAD) score and Spitzer's quality-of-life index were noted before treatment, at 1 month after treatment and every 2 months thereafter. Palliation was defined as an improvement of one point or more in the categorical scale. Equivalence was defined as less than 20% difference in the number achieving an improvement in the TSS. RESULTS: We randomised 149 patients and analysed 74 in each arm. According to the design criteria, palliation was equivalent between the two arms. TSS improved in 49 patients (77%) on regimen 1, and in 57 (92%) patients on regimen 2, a difference of 15% (95% confidence interval [CI] 3-28) in the proportion improving between the two regimens. A complete resolution of all symptoms was achieved in three (5%) on regimen 1, and in 14 (23%) patients on regimen 2 (P < 0.001), a difference in the proportion between the two regimens of 21% (95% CI 10-33). A significantly higher proportion of patients experienced palliation and complete resolution of chest pain and dyspnoea with regimen 2. No differences were observed in toxicity. The median survival was 22.7 weeks for regimen 1 and 28.3 weeks for regimen 2 (P = 0.197). CONCLUSIONS: Although this trial met the pre-determined criteria for equivalence between the two palliative regimens, significantly more patients achieved complete resolution of symptoms and palliation of chest pain and dyspnoea with the fractionated regimen.

Radical radiotherapy and salvage cystectomy as the primary management of transitional cell carcinoma of the bladder. Results following the introduction of a CT planning technique.

The objective of this study was to review the results of our policy of primary radiotherapy (RT) and salvage cystectomy for transitional carcinoma (TCC) of the bladder in the light of changes in our radiotherapy planning procedure, in particular the introduction of CT planning. The case notes of 163 patients treated with radical radiotherapy using a CT planning technique were examined. The main endpoint for assessment was response at the time of the check cystoscopy 6 months after the completion of treatment. In addition survival was estimated by stage of disease and by response at the time of first cystoscopy. Patterns of relapse and time to relapse were analysed. All percentages quoted in the text use the initial 163 patients as the denominator. One hundred patients (61%) achieved a complete response. The complete response rate was significantly related to T stage at presentation being 90% for T1, 75% for T2, and 53% for T3 disease respectively. Of these patients 78 remain disease free in the bladder (47%). Twenty-two have relapsed in the bladder, of whom 5 have also relapsed at metastatic sites. Fifteen patients have relapsed outside the bladder whilst remaining disease free within the bladder. At the time of last follow up or death from other causes 63 of the 100 patients who had a complete response remained disease free with an intact bladder. There were 18 (11%) partial responders. Seven of these patients went on to have a cystectomy. Ten remain alive, 7 disease free, 4 with intact bladders. In 24 patients (15%) there was no response and these patients have all died, the median survival being 10 months. In 21 patients (13%) a postradiotherapy cystoscopy was not performed. In all but one patient, who was lost to follow up, this was because of progressive disease. The median survival of these 20 patients was 6 months. Of the 163 patients 35% are alive and well with an intact bladder. If patients dying from other causes are included then 42% were rendered disease free. Cause specific survival was significantly related to stage of disease at presentation with 5 year actuarial survival being 87%, 48% and 26%, for T1, T2 and T3 disease respectively. Survival was also related to response to treatment at 6 months with 5 year survival being 64%, and 52% for complete and partial responders respectively. Survival was extremely poor for non-responders with only 37.5% surviving 1 year and none 5 years. There was a highly significant relationship between response and the development of, and the time to developing metastatic disease. Of those who exhibited a response 21% developed metastatic disease compared to 78% of non-responders. Salvage cystectomy offers the possibility of cure in those who achieve a complete or partial response with 42% of such patients being rendered disease free. Results however are poor in those who did not respond with all patients dying of their disease. Response rates for all stages, and survival for stages T1 and T2 are much improved from those previously reported from this centre and compare favourably with other published series. These results confirm the place of radiotherapy and salvage cystectomy in the management of TCC of the bladder in selected patients. In about one-third of patients the desired outcome of curing the patient of their cancer with organ preservation is achieved. The prognostic significance of cystoscopic response at 6 months and stage at presentation is confirmed. The outcome for patients with early stage disease is excellent. The relationship between response and the development of metastatic disease would suggest that even if these patients had had a primary cystectomy they may have fared badly, a conclusion supported by the fact that these results are comparable with surgical series. This series supports the role of radiotherapy in the management of this disease and suggests that modern RT techniques including CT planning have had a beneficial effect on the results of radical radiotherapy.

Do "conditions contributing to the death" clarify the predictors of infant mortality in Texas?

Predictors of infant mortality are usually based on the underlying causes reported on infant death certificates. Death certificates also contain space for "Other significant conditions contributing to the death but not resulting in the underlying cause." We investigated whether these "contributing conditions" might clarify the predictors of infant death in Texas or interventions that might reduce deaths from the underlying causes. A "multiple cause of death" file converted the underlying causes and contributing conditions associated with each infant death in 1996 into International Classification of Diseases codes. We ranked the major underlying causes, the conditions contributing to those causes, the conditions contributing to all infant deaths, and the effect of combining the underlying causes with contributing conditions to determine the main predictors of infant mortality. We found that extreme immaturity was the most common condition, causing or contributing to 604 of the 2081 deaths in 1996. We conclude that the underlying causes provide a misleading perception of the major predictors of infant death in Texas; combining them with the conditions contributing to those deaths provides a clearer picture. Given the powerful predictive value of preterm birth for infant mortality, the limited current ability to prevent preterm birth, and the cost of its complications, expanded efforts to improve understanding of its causal processes and develop preventive strategies should be top national and state priorities.

Long-term survival in patients with non-small cell lung cancer treated with palliative radiotherapy.

The aim of palliative thoracic radiotherapy in patients with advanced non-small cell lung cancer (NSCLC) is to alleviate symptoms. This study was designed to determine whether any patients achieved long-term survival after this treatment. In Edinburgh, between 1974 and 1993, 4531 patients were treated with palliative radiotherapy for NSCLC, receiving ten fractions or fewer. We reviewed the case notes of the long-term survivors. Sixty-one (1.3%; 95% confidence interval (CI) 1.0-1.6) patients survived for more than 5 years; 43 (70%) had histological confirmation of cancer; 28 (46%) had stage Stage I or II, 28 (46%) Stage III and one Stage IV disease; 53 (87%) patients were treated with doses of 30-35 Gy in ten daily fractions, seven (12%) received 20 Gy in five daily fractions and one received a 10 Gy single fraction. Forty-two (69%) patients had a radiological complete response, 16 (26%) a partial response and the remainder stable disease. Clinically significant radiation pneumonitis occurred in one (2%) patient, radiation myelopathy in two (3%) and multiple rib fractures in one (2%). There did not appear to be an association between long-term survival and a radiosensitive phenotype. On univariate analysis, long-term survival was more frequent in patients receiving ten-fraction regimens than in those who underwent a shorter course of radiotherapy (chi 2 = 19.5, P < 0.001). Thirty-four (0.8%; 95% CI 0.6-1.0) patients were disease free at death or at last review (median 10 years; range 5-17). We conclude that palliative thoracic radiotherapy produces long-term survival in 1.3% and personal cure in up to 1% of patients with advanced NSCLC.

Searching for preventable causes of child mortality in Texas: trends in the major causes from 1987 through 1996.

We ranked the underlying causes of child deaths in Texas from 1989 through 1991 for the groups aged 1 through 4, 5 through 9, 10 through 14, and 15 through 19 years. External causes (injuries) accounted for 66% of child deaths, and 5 accounted for at least 100 deaths each: motor vehicle accidents, homicide, suicide, drowning, and burns. Of the deaths that had potential for primary prevention, more than 95% involved accidents, suicide, and homicide. Of the smaller number of deaths that had potential for secondary prevention, treatment of infectious conditions had the greatest potential. From 1987 through 1996, child deaths from all causes averaged 2498 per year. Natural causes averaged 871 per year; external causes averaged 1627 per year. Among the external causes, annual averages for accidents were 1089 deaths; for motor vehicle accidents, 703; for homicide, 334; and for suicide, 187. We used linear regression analyses to estimate trends in deaths and mortality rates. The only categories that experienced an increased number of deaths, despite a 12% increase in the population, were deaths from all causes in the adolescent age groups; from natural causes in all but the group aged 1 through 4 years; from external causes in the adolescent age groups; from suicide in all but the group aged 1 through 4 years; and from homicide in all age groups. The increased number of deaths was often lower than the increase in the population, resulting in lower mortality rates. The only mortality rates that increased were those from all causes in the group aged 15 through 19 years; from natural causes, in both adolescent age groups; from suicide, in the group aged 10 through 14 years; and from homicide, in all but the group aged 5 through 9 years. These trends suggest that primary prevention of child deaths in Texas should focus on external causes, particularly motor vehicle accidents, homicides, and suicides.

Factors associated with fetal mortality in the triethnic population in texas, 1993 through 1995.

We examined the underlying causes of fetal deaths reported for 1993 through 1995 in relation to the mother's ethnicity, reproductive history, and medical factors listed on birth and fetal death certificates. The overall fetal mortality rate (FMR) was 6.3 per 1000 live births plus fetal deaths, and 5.6, 5.9, and 9.8 per 1000, respectively in pregnancies of white, hispanic, and black women. Preterm birth (low birth weight) was a major predictor (or consequence) of fetal death, and extremes of maternal age, increased parity, and plural pregnancies were also associated with increased FMRs, especially in black women. The underlying causes reported on fetal death certificates were complications of the placenta, cord, and membranes (36%); ill-defined conditions arising in the perinatal period (23%); fetus affected by maternal complications of pregnancy (10%); fetus affected by maternal conditions unrelated to pregnancy (6%); and short gestation and unspecified low birth weight (5%); 14% of fetal deaths were associated with congenital abnormalities. Factors reported most frequently were premature rupture of the membranes, disorders of amniotic fluid volume, pregnancy-associated hypertension, diabetes, anemia, previous preterm or small-for-gestational-age births, incompetent cervix, chronic hypertension, and sexually transmitted diseases. More than 40% of fetal deaths were not associated with any medical risk factor. FMRs associated with maternal anemia, cardiac disease, pregnancy-associated hypertension, eclampsia, and previous birth weighing more than 4 kg were at least 1.5 times higher in pregnancies of black women than in the other 2 ethnic groups. The underlying causes of fetal death provide few targets for their primary prevention; preventive efforts might better be directed to the closer surveillance and care of pregnant women with demographic, reproductive, and medical factors associated with increased risk.

Similar decreases in local tumor control are calculated for treatment protraction and for interruptions in the radiotherapy of carcinoma of the larynx in four centers.

PURPOSE: Data on patients with cancer of the larynx are analyzed using statistical models to estimate the effect of gaps in the treatment time on the local control of the tumor. METHODS AND MATERIALS: Patients from four centers, Edinburgh, Glasgow, Manchester, and Toronto, with carcinoma of the larynx and treated by radiotherapy were followed up and the disease-free period recorded. In all centers the end point was control of the primary tumor after irradiation alone. The local control rates at > or = 2 years, Pc, were analyzed by log linear models, and Cox proportional hazard models were used to model the disease-free period. RESULTS: T stage, nodal involvement, and site of the tumor were important determinants of the disease-free interval, as was the radiation schedule used. Elongation of the treatment time by 1 day, or a gap of 1 day, was associated with a decrease in Pc of 0.68% per day for Pc = 0.80, with a 95% confidence interval of (0.28, 1.08)%. An increase of 5 days was associated with a 3.5% reduction in Pc from 0.80 to 0.77. At Pc = 0.60 an increase of 5 days was associated with an 7.9% decrease in Pc. The time factor in the Linear Quadratic model, gamma/alpha, was estimated as 0.89 Gy/day, 95% confidence interval (0.35, 1.43) Gy/day. CONCLUSIONS: Any gaps (public holidays are the majority) in the treatment schedule have the same deleterious effect on the disease free period as an increase in the prescribed treatment time. For a schedule, where dose and fraction number are specified, any gap in treatment is potentially damaging.