Sir--Our study (March 14, p 801)¹ was the first to provide histological evidence that small enhancing foci on magnetic-resonance imaging (MRI) probably represents cancer foci, and it was not an extension of P J Drew and colleagues' work (May 30, p 1661)² as they suggest. Drew and colleagues seem to misunderstand the fundamental concept behind our report. To date there have been no randomised controlled studies of the impact of breast MRI on patient outcome or surgical management. The groundless assertion that breast MRI should not be deemed an experimental investigation could lead to an increase in wide local excisions or even unnecessary mastectomies. Such a change in management would most certainly be unethical since it would be based on an unproven assumption that subclinical MRI-detected cancer foci, if left surgically untreated, would lead to a worse outcome in terms of local control. Leaving the ethics aside, the evidence on the clinical significance of these additional subclinical cancer foci should be appreciated. Large studies of breast conservation have shown that more than 90% of local recurrences arise in the operated quadrant, irrespective of breast radiotherapy,³ whereas cancer foci occur throughout the breast. It is as a result of these findings that the clinical significance of cancer foci (in-situ or invasive) away from the operated quadrant has been questioned.⁴ The detection of enhancing foci by MRI, although interesting, should not lead to overzealous resections. If they do, then the lesson from large randomised studies showing the equivalence of conservative surgery over mastectomy would be ignored.

Drew and colleagues suggest that multicentric foci left behind are adequately treated by radiotherapy and tamoxifen, resulting in a recurrence rate of 8·5%, but then go on to suggest that it would be unacceptable to deliberately leave behind enhancing foci detected on MRI. These two statements seem to contradict each other and highlight the degree of uncertainty generated by the assumptions made.

To establish the natural history of breast-cancer multicentricity, C R M Boggis and co-workers (May 2, p 1362)⁵ suggest that a prospective series of patients in whom MRI is not allowed to influence patient management will answer the question we asked.¹ Although this is a feasible study, it will not tell us whether removal of the enhancing foci on MRI (probable cancer foci) is necessary. The prospective trial that we envisaged is conceptually different.

To ascertain the value of MRI in clinical management we propose randomising patients with enhancing foci to either surgical excision of these foci along with the primary tumour, or excision of the primary alone and MRI follow-up. In this way we would establish not only the natural history of enhancing foci but also whether removing these foci is feasible by conservative surgery and whether this practice would ultimately influence local recurrence rates.

*M Douek, J S Vaidya, M Baum, I Taylor

---

University College London Medical School, London W1P 7LD, UK

1 Douek M, Vaidya JS, Lakhini SR, et al. Can magnetic-resonance imaging help elucidate natural history of breast cancer multicentricity?  Lancet 1998; 351: 801­02. [Text]

2 Drew PJ, Turnbull, LW, Kerin MJ. Magnetic resonance imaging for breast cancer. Lancet 1998; 351: 1661­62. [Text]

3 Fisher ER, Anderson S, Redmond C, Fisher B. Ipsilateral breast tumour recurrence and survival following lumpectomy and irradiation: pathological findings from NSABP protocol B-06.  Semin Surg Oncol 1992; 8: 161­66. [PubMed]

4 Baum M, Vaidya JS, Mittra I. Multicentricity and recurrence of breast cancer. Lancet 1997; 349: 208.

5 Boggis CRM, Bundred NJ. Magnetic resonance imaging in breast cancer. Lancet 1998; 352: 1362.

Correspondents' reply

Sir--M Douek and colleagues¹ suggest that MRI is detecting small enhancing foci of malignant disease distant to the primary tumour. Their proposed study involves leaving these enhancing foci untreated with no further evaluation other than follow-up with MRI. In view of our results, which show that the rate of recurrence after breast conserving therapy seems to be lower than the incidence of MRI detected multifocality in patients with primary disease, the clinical relevance of these small enhancing foci remains in question.² However, no conclusion should be drawn from case series alone, and until a prospective randomised trial has shown no difference in outcome for patients treated on the basis of MRI evaluation or by standard triple assessment, we do not think that these enhancing foci can be left untreated. Suspicious mammographic lesions are always evaluated to expedite early intervention. Until proven to the contrary, we believe that lesions which are suspicious on MRI should be treated in the same way.

Furthermore, whatever the theoretical arguments about the nature of multicentricity, the proposed trial is flawed on radiological grounds. It has already been established that the inflammation and distortion caused by surgery and radiotherapy during breast conserving treatment results in a substantial reduction in the specificity of contrast-enhanced MRI for up to 18 months after treatment.³ We have an 85­90% specificity for the detection of primary disease with our MRI technique, which uses a fast dynamic sequence. Even this specificity is reduced if the scan is done too soon after breast conserving therapy.⁴ It also seems unlikely that women would accept a non-interventional wait-and-see policy, even if there was some doubt over the original MRI diagnosis. In addition, if these lesions are left in-situ, when does a clinically irrelevant enhancing focus become an invasive cancer that has an adverse prognostic effect? A more acceptable approach would be to undertake a randomised, controlled trial comparing the long-term outcome in women who have been evaluated with MRI with those who have not. Such a trial would compare the established gold standard with the new technique and would also answer some of the outstanding questions about the clinical usefulness of MRI.

MRI of the breast has now advanced to the stage at which we can relatively confidently detect multifocal disease that is undetectable by conventional imaging.² Although we accept that the technique may be oversensitive and that some of these foci may be clinically irrelevant, the fact remains that the foci are present. Until an appropriately designed prospective trial establishes whether treatment planning with MRI affects outcome, we do not feel that suspicious foci can be left in situ solely to elucidate their natural history,⁵ especially when no accurate imaging follow-up is possible in the early post-treatment phase. We cannot leave aside ethics, as Douek and colleagues suggest, because we believe that this should be one of the primary concerns of any clinical research. However, we agree that a prospective multicentre trial is needed.²

*P J Drew, L W Turnbull, M J Kerin

---

*University of Hull Academic Surgical Unit, Castle Hill Hospital, Hull HU16 5JQ, UK; and Centre for MR Investigations, Hull Royal Infirmary

1 Douek M, Vaidya JS, Lakhani SR, Hall-Craggs MA, Baum M, Taylor I. Can magnetic resonance imaging help elucidate natural history of breast cancer multicentricity.  Lancet 1998; 351: 801­02. [Text]

2 Drew PJ, Turnbull LW, Kerin MJ, Carleton PJ, Fox JN. Multicentricity and recurrence of breast cancer.  Lancet 1998; 349: 208­09. [Text]

3 Heywang-Koebrunner SH, Schlegel A, Beck R, et al. Contrast enhanced MRI of the breast after limited surgery and radiation therapy.  JCAT 1993; 17: 891­900. [PubMed]

4 Drew PJ, Kerin MJ, Turnbull LW, et al. Routine screening for local recurrence following breast conserving therapy for cancer with dynamic contrast enhanced magnetic resonance imaging of the breast.  Ann Surg Onc 1998; 5: 265­70. [PubMed]

5 Drew PJ, Turnbull LW, Kerin MJ. Magnetic-resonance imaging for breast cancer.  Lancet 1998; 351: 1661­62. [Text]