1992, The British Journal of Radiology, 65, 30-32
Chest radiography in the management of breast cancer By E. Moskovic, MRCP, FRCR, C. Parsons, FRCS, FRCR and * M . Baum, C h M , FRCS Departments of Radiology and 'Academic Surgery, Royal Marsden Hospital, Fulham Road, London SW3 6JJ, UK
{Received 20 May 1991 and in revised form 3 July 1991, accepted 23 July 1991) Keywords: Breast cancer, Chest radiograph, Metastasis, Imaging
Abstract. This study was undertaken to assess the utility of chest radiography (CXR) in the management of patients with breast cancer and to devise a suitable imaging policy for such patients. A retrospective analysis was performed of the case notes and the CXRs of a series of 141 patients presenting to this hospital in 1980 with a diagnosis of early breast cancer. Data retrieved covered the entire clinical course to date and included the number of CXRs performed for a clinical reason and the number performed as "routine". Correlation of the result of every CXR requested to subsequent therapeutic decision-making throughout the complete clinical course of each patient was undertaken and the proportion of CXRs initiating a management change, or performed for a clinical reason, out of the total requested was assessed. A total of 1161 CXRs were performed on the 141 patients studied. Of these, only 174 (15%) were undertaken for a direct clinical reason and 987 (85%) were undertaken as part of "routine" follow-up procedures and had no impact on patient management. Four "routine" CXRs ( < 0.4%) demonstrated previously undiagnosed pulmonary metastases, in patients with no other history of metastatic disease. Thus, routine chest radiography is not a costeffective method of monitoring asymptomatic patients with breast cancer for metastasis. It is recommended that outside staging procedures for clinical trials, the CXR is used only to address a clinical problem relating specifically to the thorax. It is anticipated that the cessation of all other chest radiography in patients being followed up for breast cancer at this institution will have significant budgetary implications.
There is little evidence that the finding of hitherto undiagnosed pulmonary metastases on a routine "follow-up" chest radiograph (CXR) improves the current outlook for the patient with early breast cancer (Horton, 1984; Dewar & Kerr, 1985; Ciatto et al, 1989; Rutgers et al, 1989). Indeed, despite the large battery of screening tests available to the clinician to monitor such patients (radiological, biochemical and otherwise), no survival benefit is conferred by the detection of any clinically occult metastatic disease and many workers now feel that tests on these patients should only be undertaken for new symptoms or clinical findings (Zwaveling et al, 1987; Mansi et al, 1988; Kindler & Steinhoff, 1989; Rutgers et al, 1989). The current emphasis on resource management, particularly in the field of oncology, demands rationalizing expenditure so that only the tests which contribute directly to patient care and outcome are employed, thereby allowing finite reserves to be directed most efficiently. With this in mind, the following study was undertaken to identify a role for chest radiography in the current management of breast cancer at all stages of the disease. Methods
The case notes of a series of 141 patients presenting to the Royal Marsden Hospital, London during 1980 with early (non-metastatic) breast cancer were reviewed. Address correspondence to Dr Eleanor Moskovic, Senior Lecturer, Department of Radiology, Royal Marsden Hospital, Fulham Road, London, SW3 6JJ.
30
Primary treatment with curative intent included mastectomy (75 patients) and local excision with axillary clearance (66 patients). 130 patients received post-operative irradiation depending on clinicopathological findings. The following information concerning all CXRs performed throughout the subsequent clinical course of each patient was retrieved: (i) date and clinical reason given for CXR request, (ii) radiologist's report of CXR, i.e. normal/abnormal (if abnormal: change/no change from previous examination), (iii) total number of CXR examinations performed. The outcome for each patient was documented, with follow-up to date (1990) whenever possible. The number of CXRs performed for a clinical reason in all patients was obtained and expressed as a fraction of the total. The pattern of CXR requests and the impact of the result on therapeutic decision-making was evaluated both for early and advanced disease. Results
Of the 141 patients reviewed 50 (35%) subsequently developed recurrent or metastatic disease from which 43 patients have since died; 2 have died from other causes and 89 (63%) are alive and well at the last follow-up visit (mean follow-up period 8.7 years). A total of 1161 CXRs were performed on 141 patients (Table la). Of these, 174 (15%) were requested either for new clinical symptoms or for a clinically-directed management problem (Table II) which relied on the CXR result. 987 (85%) CXRs were undertaken either as The British Journal of Radiology, January 1992
Chest radiography in the management of breast cancer Table 1 (a). Overall results Number of patients studied 141 Total number of CXRs performed 1161 Number of CXRs performed for a clinical reason (see Table II) 174(15%) Number of CXRs performed as "routine" follow-up, or for unidentifiable reason 987 (85%)
Table I (b). Number of CXRs undertaken according to patient outcome Outcome
No recurrence/metastasis: (mean FU = 8.7 years) Local recurrence only Distant metastasis
Number of patients (total =141)
91 7 43
Number of CXRs
Number performed for a clinical reason
(%)
634 72 455
14 9 151
2.2% 12% 33%
Table I (c). Number of CXRs undertaken according to site of first metastasis Site of first metastasis
Number of patients (total = 43)
Number of CXRs
Number performed for a clinical reason
(%)
Lung/pleura Bone Liver
15 18 10
189 164 102
81 45 25
43% 27% 24%
part of a "routine" yearly follow-up procedure, or for unknown reasons not identifiable from the case notes and not resulting in a management change. In four patients asymptomatic pulmonary metastases were identified on "routine" follow-up CXRs. This represents an asymptomatic yield in less than 3% of the total number of patients examined and less than 0.4% of the total number of "routine" CXRs performed. All four of these patients subsequently developed symptomatic metastatic disease (either pulmonary or bony) within 6 months (range 3-6 months) of the abnormal "routine" CXR and all have since died of their disease. The group of 91 patients with no evidence of disease at last attendance had a total of 634 CXRs performed during follow-up (Table Ib), of which 14 (2%) were Table II. Clinical reasons for CXR requests (i) Thoracic symptoms: dyspnoea, pleuritic or rib pain, cough, chest infection, cardiac failure, (ii) Monitoring response to chemotherapy or other treatment, (iii) Monitoring interventional procedures: placement of chest drains, central lines, Hickman lines, management of pneumothorax. (iv) Pre-operative. (v) Investigation of other disorders, e.g. iron-deficiency anaemia.
Vol. 65, No. 769
performed for clinical reasons as described in Table II. Seven patients developed local recurrent disease (median time to recurrence: 21 months) and were treated successfully with surgery with or without radiotherapy; they are all well to date. A total of 72 CXRs were undertaken in this group, nine (12%) of these were performed for a clinical reason (usually pre-operative) and the remaining 63 (87%) were "routine" follow-up films. 43 patients developed distant metastatic disease and have since died. The median time from initial diagnosis to metastasis was 24 months, with median time to death being 38 months. The commonest site for initial metastasis was bone (18 patients), followed by lung and/or pleura (15 patients) and then liver (10 patients). A total of 455 CXRs were performed on these patients of these 151 (33%) were undertaken for clinical reasons (Table Ic) and 304 (66%) were performed either for reasons undocumented or for "routine" follow-up. Of the patients in whom the lung was the site of first metastasis, CXRs were performed with the greatest frequency and also with the greatest frequency of clinical indication. Discussion
Overall, only 15% of all the CXRs performed on patients reviewed in this study contributed directly to their subsequent clinical management. Of the 91 patients 31
E. Moskovic, C. Parsons and M. Baum
followed up with no evidence of either local recurrence or distant metastasis, only 2% of CXR requests were undertaken for a clinical reason and only 33% of CXRs performed on patients who developed metastatic disease were shown to be of clinical value. At this institution, the X-ray department performs approximately 2500 CXRs annually on patients with breast cancer at all stages of the disease. The yearly total of CXRs performed for all diseases in this department is around 9000, so imaging patients with breast cancer represents nearly one third of the overall CXR workload. The CXR is the single most frequent investigation requested at the Royal Marsden Hospital X-ray department, constituting over 60% of all plain film examinations undertaken. Although early detection of local or regional recurrent breast cancer is undoubtedly worthwhile, there is much data to suggest that screening for asymptomatic metastatic disease in the follow-up of breast cancer is neither cost-effective nor beneficial to survival (Logager et al, 1990). Most studies, including this one, show a positive yield of less than 4%, both for routine chest radiography and for other tests, and the lead time conferred by the finding of such disease has not been clinically advantageous (Dewar & Kerr, 1985; Hughes & Courtney, 1985; Ciatto et al, 1989; Vestegaard et al, 1989). Many groups now advocate careful history taking and clinical examination for both the evaluation of local disease and distant relapse, with emphasis on flexible out-patient attendance depending on patient symptomatology, at which time only relevant tests are performed. However, the practice of standardized 6 monthly or yearly out-patient attendance involving a variety of investigations appears hard to change, perhaps because asymptomatic follow-up patients are often reviewed by relatively junior medical staff who may simply repeat the management policy of their predecessor. The pattern of CXR requests in patients with known metastatic disease is also haphazard. This study has found that even in cases with known thoracic disease, only 43% of CXRs contributed usefully to therapeutic decision-making and of all patients who died of metastatic breast cancer, clinical reasons for CXR requests could only be found in one third of cases. In many instances, CXR requests on such patients were combined with requests for other radiographs, such as spinal and pelvic examinations during management of bony disease, for which the CXR result was irrelevant. Clinical reasons for requesting CXRs in this study (Table II) include the pre-operative chest film, which was standard practice in this hospital in 1980 when this group of patients first presented. This practice has been rationalized over recent years in line with other units and the current policy at this institution is to perform such examinations on patients over the age of 50, or with a specific cardiac or pulmonary disorder justifying further investigation. The cost-effectiveness of performing routine pre-operative CXRs on all patients has been shown conclusively to be poor (Roberts et al, 32
1985). Its use as a staging procedure in the asymptomatic patient with early breast cancer is more controversial, but the positive yield in this circumstance is again extremely low (Feig, 1987). By carefully reviewing all imaging policies, not only in breast cancer, so that only the examinations that contribute directly to patient management are performed, significant savings may be made within a financially constrained health service. Following this study, it is recommended that routine CXRs are not performed on patients being followed up after treatment for primary breast cancer at this institution. Chest radiography should be used to address specific clinical problems relating to the thorax, both in patients with early and advanced disease. It is estimated that reducing the number of unnecessary CXRs performed in this department on such patients will cut the overall CXR workload by over 2000 examinations annually, resulting in inevitable savings of both valuable budgetary resources and unnecessary irradiation to both patients and staff (NRPB, 1990). References CIATTO, S., PACINI, P., ANDREOLI, C , CECCHINI, S., IOSSA, A.,
GRAZZINI, G., 1989. Chest X-ray in the follow-up of breast cancer patients. British Journal of Cancer, 60, 102-103. DEWAR, J. A. & KERR, G. R., 1985. Value of routine follow-up of women treated for early carcinoma of the breast. British Medical Journal, 291, 1464-1467. FEIG, S. A., 1987. Imaging techniques and guidelines for evaluation and follow-up of breast cancer patients. Critical Reviews in Diagnostic Imaging, 27 (1), 1-15. HORTON, J., 1984. Follow-up of breast cancer patients. Cancer, 53, 790-797. HUGHES, L. E. & COURTNEY, S. P., 1985. Follow-up of patients
with breast cancer. British Medical Journal, 290, 1229-1230. KINDLER, M. & STEINHOFF, G., 1989. Follow-up of breast
cancer patients. Oncology, 46, 360-365. LOGAGER, V. B., VESTERGAARD, A., HERRSTEDT, J., THOMSEN, H. S., ZEDELER, K. & DOMBERNOWSKY, P., 1990. The limited
value of routine chest X-ray in the follow-up of Stage II breast cancer. European Journal of Cancer, 26 (5), 553-555. MANSI, J. L., EARL, H. M., POWLES, T. J. & COOMBES, R. C ,
1988. Tests for detecting recurrent disease in the follow-up of patients with breast cancer. Breast Cancer Research and Treatment, 11, 249-254. NRPB, 1990. Patient Dose Reduction In Diagnostic Radiology. Documents of the NRPB, Vol. 1 (3) (HMSO, London). ROBERTS, C. J., FARROW, S. C. & CHARNY, M. C , 1985. How
much can the NHS afford to spend to save a life or avoid a severe disability? Lancet, 1 (8420), 89-91. RUTGERS, E. J. T H . , VAN SLOOTEN, E. A. & KLUCK, 1989.
Follow-up after treatment of primary breast cancer. British Journal of Surgery, 76, 187-190. VESTERGAARD, A., HERRSTEDT, J., THOMSEN, H. S., DOMBERNOWSKY, P. & ZEDELER, K., 1989. The value of
yearly chest X-ray in patients with Stage I breast cancer. European Journal of Cancer Clinical Oncology, 25 (4), 687-689. ZWAVELING, A., ALBERS, G. H. R., FELTHUIS, W. & HERMANS,
J., 1987. A evaluation of routine follow-up for detection of breast cancer recurrencies. Journal of Surgical Oncology 34, 194-197. The British Journal of Radiology, January 1992
Chest radiography in the management of breast cancer By E. Moskovic, MRCP, FRCR, C. Parsons, FRCS, FRCR and * M . Baum, C h M , FRCS Departments of Radiology and 'Academic Surgery, Royal Marsden Hospital, Fulham Road, London SW3 6JJ, UK
{Received 20 May 1991 and in revised form 3 July 1991, accepted 23 July 1991) Keywords: Breast cancer, Chest radiograph, Metastasis, Imaging
Abstract. This study was undertaken to assess the utility of chest radiography (CXR) in the management of patients with breast cancer and to devise a suitable imaging policy for such patients. A retrospective analysis was performed of the case notes and the CXRs of a series of 141 patients presenting to this hospital in 1980 with a diagnosis of early breast cancer. Data retrieved covered the entire clinical course to date and included the number of CXRs performed for a clinical reason and the number performed as "routine". Correlation of the result of every CXR requested to subsequent therapeutic decision-making throughout the complete clinical course of each patient was undertaken and the proportion of CXRs initiating a management change, or performed for a clinical reason, out of the total requested was assessed. A total of 1161 CXRs were performed on the 141 patients studied. Of these, only 174 (15%) were undertaken for a direct clinical reason and 987 (85%) were undertaken as part of "routine" follow-up procedures and had no impact on patient management. Four "routine" CXRs ( < 0.4%) demonstrated previously undiagnosed pulmonary metastases, in patients with no other history of metastatic disease. Thus, routine chest radiography is not a costeffective method of monitoring asymptomatic patients with breast cancer for metastasis. It is recommended that outside staging procedures for clinical trials, the CXR is used only to address a clinical problem relating specifically to the thorax. It is anticipated that the cessation of all other chest radiography in patients being followed up for breast cancer at this institution will have significant budgetary implications.
There is little evidence that the finding of hitherto undiagnosed pulmonary metastases on a routine "follow-up" chest radiograph (CXR) improves the current outlook for the patient with early breast cancer (Horton, 1984; Dewar & Kerr, 1985; Ciatto et al, 1989; Rutgers et al, 1989). Indeed, despite the large battery of screening tests available to the clinician to monitor such patients (radiological, biochemical and otherwise), no survival benefit is conferred by the detection of any clinically occult metastatic disease and many workers now feel that tests on these patients should only be undertaken for new symptoms or clinical findings (Zwaveling et al, 1987; Mansi et al, 1988; Kindler & Steinhoff, 1989; Rutgers et al, 1989). The current emphasis on resource management, particularly in the field of oncology, demands rationalizing expenditure so that only the tests which contribute directly to patient care and outcome are employed, thereby allowing finite reserves to be directed most efficiently. With this in mind, the following study was undertaken to identify a role for chest radiography in the current management of breast cancer at all stages of the disease. Methods
The case notes of a series of 141 patients presenting to the Royal Marsden Hospital, London during 1980 with early (non-metastatic) breast cancer were reviewed. Address correspondence to Dr Eleanor Moskovic, Senior Lecturer, Department of Radiology, Royal Marsden Hospital, Fulham Road, London, SW3 6JJ.
30
Primary treatment with curative intent included mastectomy (75 patients) and local excision with axillary clearance (66 patients). 130 patients received post-operative irradiation depending on clinicopathological findings. The following information concerning all CXRs performed throughout the subsequent clinical course of each patient was retrieved: (i) date and clinical reason given for CXR request, (ii) radiologist's report of CXR, i.e. normal/abnormal (if abnormal: change/no change from previous examination), (iii) total number of CXR examinations performed. The outcome for each patient was documented, with follow-up to date (1990) whenever possible. The number of CXRs performed for a clinical reason in all patients was obtained and expressed as a fraction of the total. The pattern of CXR requests and the impact of the result on therapeutic decision-making was evaluated both for early and advanced disease. Results
Of the 141 patients reviewed 50 (35%) subsequently developed recurrent or metastatic disease from which 43 patients have since died; 2 have died from other causes and 89 (63%) are alive and well at the last follow-up visit (mean follow-up period 8.7 years). A total of 1161 CXRs were performed on 141 patients (Table la). Of these, 174 (15%) were requested either for new clinical symptoms or for a clinically-directed management problem (Table II) which relied on the CXR result. 987 (85%) CXRs were undertaken either as The British Journal of Radiology, January 1992
Chest radiography in the management of breast cancer Table 1 (a). Overall results Number of patients studied 141 Total number of CXRs performed 1161 Number of CXRs performed for a clinical reason (see Table II) 174(15%) Number of CXRs performed as "routine" follow-up, or for unidentifiable reason 987 (85%)
Table I (b). Number of CXRs undertaken according to patient outcome Outcome
No recurrence/metastasis: (mean FU = 8.7 years) Local recurrence only Distant metastasis
Number of patients (total =141)
91 7 43
Number of CXRs
Number performed for a clinical reason
(%)
634 72 455
14 9 151
2.2% 12% 33%
Table I (c). Number of CXRs undertaken according to site of first metastasis Site of first metastasis
Number of patients (total = 43)
Number of CXRs
Number performed for a clinical reason
(%)
Lung/pleura Bone Liver
15 18 10
189 164 102
81 45 25
43% 27% 24%
part of a "routine" yearly follow-up procedure, or for unknown reasons not identifiable from the case notes and not resulting in a management change. In four patients asymptomatic pulmonary metastases were identified on "routine" follow-up CXRs. This represents an asymptomatic yield in less than 3% of the total number of patients examined and less than 0.4% of the total number of "routine" CXRs performed. All four of these patients subsequently developed symptomatic metastatic disease (either pulmonary or bony) within 6 months (range 3-6 months) of the abnormal "routine" CXR and all have since died of their disease. The group of 91 patients with no evidence of disease at last attendance had a total of 634 CXRs performed during follow-up (Table Ib), of which 14 (2%) were Table II. Clinical reasons for CXR requests (i) Thoracic symptoms: dyspnoea, pleuritic or rib pain, cough, chest infection, cardiac failure, (ii) Monitoring response to chemotherapy or other treatment, (iii) Monitoring interventional procedures: placement of chest drains, central lines, Hickman lines, management of pneumothorax. (iv) Pre-operative. (v) Investigation of other disorders, e.g. iron-deficiency anaemia.
Vol. 65, No. 769
performed for clinical reasons as described in Table II. Seven patients developed local recurrent disease (median time to recurrence: 21 months) and were treated successfully with surgery with or without radiotherapy; they are all well to date. A total of 72 CXRs were undertaken in this group, nine (12%) of these were performed for a clinical reason (usually pre-operative) and the remaining 63 (87%) were "routine" follow-up films. 43 patients developed distant metastatic disease and have since died. The median time from initial diagnosis to metastasis was 24 months, with median time to death being 38 months. The commonest site for initial metastasis was bone (18 patients), followed by lung and/or pleura (15 patients) and then liver (10 patients). A total of 455 CXRs were performed on these patients of these 151 (33%) were undertaken for clinical reasons (Table Ic) and 304 (66%) were performed either for reasons undocumented or for "routine" follow-up. Of the patients in whom the lung was the site of first metastasis, CXRs were performed with the greatest frequency and also with the greatest frequency of clinical indication. Discussion
Overall, only 15% of all the CXRs performed on patients reviewed in this study contributed directly to their subsequent clinical management. Of the 91 patients 31
E. Moskovic, C. Parsons and M. Baum
followed up with no evidence of either local recurrence or distant metastasis, only 2% of CXR requests were undertaken for a clinical reason and only 33% of CXRs performed on patients who developed metastatic disease were shown to be of clinical value. At this institution, the X-ray department performs approximately 2500 CXRs annually on patients with breast cancer at all stages of the disease. The yearly total of CXRs performed for all diseases in this department is around 9000, so imaging patients with breast cancer represents nearly one third of the overall CXR workload. The CXR is the single most frequent investigation requested at the Royal Marsden Hospital X-ray department, constituting over 60% of all plain film examinations undertaken. Although early detection of local or regional recurrent breast cancer is undoubtedly worthwhile, there is much data to suggest that screening for asymptomatic metastatic disease in the follow-up of breast cancer is neither cost-effective nor beneficial to survival (Logager et al, 1990). Most studies, including this one, show a positive yield of less than 4%, both for routine chest radiography and for other tests, and the lead time conferred by the finding of such disease has not been clinically advantageous (Dewar & Kerr, 1985; Hughes & Courtney, 1985; Ciatto et al, 1989; Vestegaard et al, 1989). Many groups now advocate careful history taking and clinical examination for both the evaluation of local disease and distant relapse, with emphasis on flexible out-patient attendance depending on patient symptomatology, at which time only relevant tests are performed. However, the practice of standardized 6 monthly or yearly out-patient attendance involving a variety of investigations appears hard to change, perhaps because asymptomatic follow-up patients are often reviewed by relatively junior medical staff who may simply repeat the management policy of their predecessor. The pattern of CXR requests in patients with known metastatic disease is also haphazard. This study has found that even in cases with known thoracic disease, only 43% of CXRs contributed usefully to therapeutic decision-making and of all patients who died of metastatic breast cancer, clinical reasons for CXR requests could only be found in one third of cases. In many instances, CXR requests on such patients were combined with requests for other radiographs, such as spinal and pelvic examinations during management of bony disease, for which the CXR result was irrelevant. Clinical reasons for requesting CXRs in this study (Table II) include the pre-operative chest film, which was standard practice in this hospital in 1980 when this group of patients first presented. This practice has been rationalized over recent years in line with other units and the current policy at this institution is to perform such examinations on patients over the age of 50, or with a specific cardiac or pulmonary disorder justifying further investigation. The cost-effectiveness of performing routine pre-operative CXRs on all patients has been shown conclusively to be poor (Roberts et al, 32
1985). Its use as a staging procedure in the asymptomatic patient with early breast cancer is more controversial, but the positive yield in this circumstance is again extremely low (Feig, 1987). By carefully reviewing all imaging policies, not only in breast cancer, so that only the examinations that contribute directly to patient management are performed, significant savings may be made within a financially constrained health service. Following this study, it is recommended that routine CXRs are not performed on patients being followed up after treatment for primary breast cancer at this institution. Chest radiography should be used to address specific clinical problems relating to the thorax, both in patients with early and advanced disease. It is estimated that reducing the number of unnecessary CXRs performed in this department on such patients will cut the overall CXR workload by over 2000 examinations annually, resulting in inevitable savings of both valuable budgetary resources and unnecessary irradiation to both patients and staff (NRPB, 1990). References CIATTO, S., PACINI, P., ANDREOLI, C , CECCHINI, S., IOSSA, A.,
GRAZZINI, G., 1989. Chest X-ray in the follow-up of breast cancer patients. British Journal of Cancer, 60, 102-103. DEWAR, J. A. & KERR, G. R., 1985. Value of routine follow-up of women treated for early carcinoma of the breast. British Medical Journal, 291, 1464-1467. FEIG, S. A., 1987. Imaging techniques and guidelines for evaluation and follow-up of breast cancer patients. Critical Reviews in Diagnostic Imaging, 27 (1), 1-15. HORTON, J., 1984. Follow-up of breast cancer patients. Cancer, 53, 790-797. HUGHES, L. E. & COURTNEY, S. P., 1985. Follow-up of patients
with breast cancer. British Medical Journal, 290, 1229-1230. KINDLER, M. & STEINHOFF, G., 1989. Follow-up of breast
cancer patients. Oncology, 46, 360-365. LOGAGER, V. B., VESTERGAARD, A., HERRSTEDT, J., THOMSEN, H. S., ZEDELER, K. & DOMBERNOWSKY, P., 1990. The limited
value of routine chest X-ray in the follow-up of Stage II breast cancer. European Journal of Cancer, 26 (5), 553-555. MANSI, J. L., EARL, H. M., POWLES, T. J. & COOMBES, R. C ,
1988. Tests for detecting recurrent disease in the follow-up of patients with breast cancer. Breast Cancer Research and Treatment, 11, 249-254. NRPB, 1990. Patient Dose Reduction In Diagnostic Radiology. Documents of the NRPB, Vol. 1 (3) (HMSO, London). ROBERTS, C. J., FARROW, S. C. & CHARNY, M. C , 1985. How
much can the NHS afford to spend to save a life or avoid a severe disability? Lancet, 1 (8420), 89-91. RUTGERS, E. J. T H . , VAN SLOOTEN, E. A. & KLUCK, 1989.
Follow-up after treatment of primary breast cancer. British Journal of Surgery, 76, 187-190. VESTERGAARD, A., HERRSTEDT, J., THOMSEN, H. S., DOMBERNOWSKY, P. & ZEDELER, K., 1989. The value of
yearly chest X-ray in patients with Stage I breast cancer. European Journal of Cancer Clinical Oncology, 25 (4), 687-689. ZWAVELING, A., ALBERS, G. H. R., FELTHUIS, W. & HERMANS,
J., 1987. A evaluation of routine follow-up for detection of breast cancer recurrencies. Journal of Surgical Oncology 34, 194-197. The British Journal of Radiology, January 1992
