Internal mammary lymphoscintigraphy in the conservative management of breast carcinoma: An update and recommendations for a new TNM staging

ChntcalRa&ology(1985)36, 469-472 © 1985Royal College of Radiologists

0009-9260/85/497469502 00

Internal Mammary Lymphoscintigraphy in the Conservative Management of Breast Carcinoma" An Update and Recommendations for a New TNM Staging GI~INE~ N. EGE and R. M. CLARK*

Departments of Nuclear Medicine and *Radiation Oncology, The Princess Margaret Hospital, Toronto, Ontario, Canada

The results of internal mammary lymphoscintigraphy (IML) in 524 patients with breast carcinoma who underwent partial mastectomy, with or without axillary dissection, demonstrate the predictive value of an abnormal lymphoscintigram. There was a significant difference (P <0.0005) in actuarial survival between patients with normal and abnormal IML. In view of the increased mortality associated with metastases to both internal mammary and axillary lymphatics compared with involvement of either site alone, and the most favourable outcome in patients with involvement of neither site, it is proposed that the current TNM staging for breast carcinoma should henceforth include the status of internal mammary lymph nodes determined by radiocolloid lymphoscintigraphy. Further refinement of the N status into N A (axilla) and N I (internal mammary nodes) is suggested. Stages I, II, and IIl should designate, respectively, patients with no evidence of involvement of either lymphatic site, patients with involvement of one or the other site and patients with involvement of both sites. This would constitute a more rational approach to the staging and management of the patient with breast carcinoma and would also provide a more accurate means of evaluating the true efficacy of current strategies in comparable groups of patients. Unrecognised internal mammary lymphatic metastases have probably confounded the results of past prospective trials.

Many tumour markers for breast carcinoma have been identified and used for prognostic purposes but the extent of involvement of the axillary lymph nodes with metastatic disease remains the most sensitive indicator. Although surgical experience has demonstrated that patients with internal mammary lymph node involvement have a uniformly unfavourable outcome (Bucalossi et al., 1971; Haagenson, 1971; Urban and Marjani, 1971; Livingston and Arlen, 1974; Handley, 1975; Veronesi and Valagussa, 1981), this fact is largely ignored in current staging systems and in the design of prospective studies to evaluate adjuvant therapy. Given the implications of lymph node involvement, it is unlikely that any major decrease in the morbidity and mortality of breast carcinoma can be achieved until management strategy takes into account the status of both, and not just one, of the major lymphatic drainage pathways of the breast. Correspondence to: Dr Gun% N. Ege, Department of Nuclear Medicine, The Princess Margaret Hospital, 500 Sherbourne Street, Toronto, Ontario M4X IK9, Canada.

In 198(/ we reported the prognostic value of radiocolloid internal mammary lymphoscintigraphy (IML) in 249 women who had undergone a conservative surgical procedure (partial mastectomy), with or without axillary sampling, between January 1974 and December 1977 (Ege and Clark, 1980). Accrual has continued and we now report our findings in a larger group of patients. Based on this experience and other data on IML available to date, we propose that consideration be given to a new TNM staging system for breast carcinoma. PATIENTS AND METHODS Between January 1974 and December 1980, 778 women with breast carcinoma who had undergone partial mastectomy (excisional biopsy, lumpectomy or wedge resection) were referred to the Princess Margaret Hospital, Toronto. All patients in this series had clinically negative axillae and the policy pursued precluded an axillary dissection in the majority of the patients. Internal mammary lymphoscintigraphy was part of the initial assessment of 524 (67%) of the patients. The investigation was carried out according to the established protocol of ipsilateral and contralateral subcostal posterior rectus sheath injections of 20 MBq (500 /2Ci) of 99mTc antimony sulphide colloid (99mTcSb2S3) per side. Images were obtained after 3 h and 6 h, with a large-field-of-view gamma camera. interpretation was carried out according to established criteria (Ege, 1976, 1983). The actuarial life-table method was used to estimate the probability of survival from diagnosis, in order to take full advantage of partial information available on patients under observation for less than the designated time. The Wilcoxon-Gehan test was used to test for significant differences in actuarial survival between patients with normal, suspicious and abnormal IML. RESULTS The groups of patients who did and those who did not undergo IML were comparable; 64% and 70%, respectively, were post-menopausal. Seventy per cent of patients in both groups had pathological T~ or T 2 lesions and 20%T x lesions. Axillary lymph node sampling was carried out in 158 (20%) of the 778 patients, being performed in 18% of patients who had IML and 25% of those who did not (Table 1).

470

CLINICAL RADIOLOGY

higher than results published by Ege and Elhakim (1984), whose study population did not include any axillary node positive patients. As illustrated in Table 3, there was no significant difference in menopausal status, pathological T stage and pathological N stage between patient subgroups according to IML results. This excludes the possibility of a preponderance of unfavourable prognostic factors, which may have been independently predictive, amongst patients with an abnormal lymphoscintigram. Table 4 illustrates the distribution of patients receiving radiotherapy. No other adjuvant therapy was administered. The clinical status of patients after a minimum follow-up period of 2 years, comparing those who had IML with those who did not, is shown in Table 5, with the actuarial surival illustrated in Fig. 1.

Table 1 - Axillary status

Pathological N stage

No of pattents

No NI Nx (no @ssecaon) Total

IML

No 1ML

Total

57 (11%) 39 (7%) 428 (82%)

32 (13%) 30 (12%) 192 (75%)

89 69 620

524 (100%)

254 (100%)

778

Table 2 - Results of IML in 524 patients

IML

No. of patients

%

Normal Suspicious Abnormal Equivocal Total

349 70 87 18 524

67 13 17 3 100

Table 5 - Clinical status of patients after a m i n i m u m follow-up of 2 years

The results of IML carried out on 524 patients are shown in Table 2. A 17% incidence of abnormal lymphoscintigrams in this group is identical to our earlier report on 249 patients (Ege and Clark, 1980) and only slightly I

I

I

[

80

I

..

IML result

Total DtseaseNo. of free pattents

Normal Suspicious Abnormal Patmnts who had IML Patients who had no IML

(349) (70) (87)

Local Dtstant recurrence disease

Overall recurrence

239 (69%) 78 (22%) 37 (53%) 20 (28%) 46 (53%) 23 (26%)

32 (9%) 110 (31%) 13 (19%) 33 (47%) 18 (21%) 41 (47%)

160 (63%) 56 (22%)

38 (15%) 98 (37%)

506* 254

* Excludes 18 patients with equivocal IML K

60

u~

Normolvs Abnormal NormaLvs Suspicious

"~

Abnormal

"--" ....

p<00005

DISCUSSION

p
vs Susplclous p : 0 . 6 0 2

4o "~-

--Normol ...... Suspm~ous

'~ 20

--- Abnormal -S c a n n G dane

0

I

I

I

l

I

I

2

5

4

5

Years since diagnosis

5~N

Fig. 1 - Actuarial survival of 760 patients who underwent partial mastectomy with or without axillary dissection. Of 506 patients who had IML, 349 had a normal, 70 suspmious and 87 an abnormal study. Eighteen patients with an equivocal study were excluded; 254 patients did not have IML carried out.

Despite an extensive body of literature, there are some who still maintain that internal mammary lymphoscintigraphy has little credibility until supported by an extensive surgical study (Shibata, 1983). However, technical and ethical considerations preclude such a correlative study beyond that which has already been reported. Matsuo (1974) has carried out preoperative IML on 106 patients who subsequently underwent extended radical mastectomy. Histological correlation yielded a sensitivity of 1.0 and a specificity of 0.96. Although the occasional biopsy of internal mammary lymph nodes in a patient with abnormal IML may

Table 3 - Distribution within patient subgroups, according to IML results, of menopausal status, pathological T stage and pathological N stage*

IML result

Normal (349) Suspicious (70) Abnormal (87) Equivocal (18)

Menopausal status

Pathological T stage

Pre-

Post-

T:

T:

Tx

Nu

Pathological N stage N:

Nx

135 (39%) 25 (36%) 26 (30%) 4 (14%)

214 (61%) 45 (64%) 61 (70%) 14 (78%)

172 (49%) 33 (47%) 41 (47%) 10 (56%)

77 (22%) 12 (17%) 16 (18%) 5 (26%)

70 (20%) 18 (26%) 23 (26%) 1 (6%)

44 (13%) 7 (10%) 4 (5%) 2 (11%)

26 (7%) 4 (6%) 7 (8%) 0

279 (80%i 59 (84%) 74 (84%) 16 (89%)

* Excludes: m-situ, 5; N2,3, 2; T~ 11; bilateral, 4; multlfocal, 26. Table 4 - Distribution of patients receiving no further treatment or post-operative radiotherapy (RT)

IML result

Total No. of patients

No RT

R T to breast only

RT to breasts and nodes

R T to nodes only

Normal Suspicious Abnormal Equivocal No IML Total

349 (100%) 70 (100%) 87 (100%) 18 (100%) 254 (100%) 778 (100%)

77 (22%) 24 (34%) 16 (18%) 4 (22%) 106 (42%) 227 (29%)

150 (43%) 18 (26%) 18 (21%) 7 (39%) 7 4 (29%)

114 (33%) 23 (3%) 51 (59%) 6 (33%) 71 (28%) 551~(71%)

8 (2%) 5 (7%) 2 (2%) 1 (6%) 3 (1%),

LYMPHOSCINTIGRAPHY IN TNM STAGING

be feasible, the surgical sample obtained may not reflect the status of the entire parasternal lymphatic network. Furthermore, there are few who would undertake systematic biopsy of lymphatics assessed to be normal by IML. The insistence on further histopathological correlation is, therefore, unlikely to bear fruit and the accuracy and credibility of IML must rest on alternative evidence. Hill etal. (1983) reported a study in which 121 patients with breast lumps underwent bilateral IML and axillary lymphoscintigraphy prior to biopsy. Of these patients, 76 had benign disease and 45 had cancer proven by biopsy. Of the 76 patients with benign disease, 75 had IML interpreted as normal by three independent observers, yielding a specificity of 0.98. Of the 44 patients with cancer whose lymphoscintigrams were technically satisfactory, 10 (23%) had abnormal IML and seven of these had histologically proven axillary metastases. The remaining three patients, whose axillae were negative, had medial and central lesions. Biopsy was carried out on internal mammary nodes in two patients and confirmed the pre-operative diagnosis by IML in both instances. Ege and Elhakim (1984) have reviewed retrospectively 981 patients with primary breast carcinoma, looking specifically at the implications of the results of IML with histologically and clinically negative axillae, irrespective of turnout site or size. An incidence of abnormal IML of 14% is consistent with that expected in a group of patients with no axillary nodal disease (Bucatossi et al., 1971; Haagenson, 1971; Urban and Marjani, 1971; Livingston and Arlen, 1974; Handley, 1975; Veronesi and Valagussa, 1981). At the end of a 3-year follow-up there was a statistically significant (P~<0.005) increase in the rate of local and distant relapse in patients with abnormal IML. From Fig. 1 it is clear that two prognostically dissimilar groups of patients can be distinguished by IML results in the present study. As in the previous study (Ege and Clark, 1980), the difference in actuarial survival between patients with normal and abnormal IML is statistically significant (P<0.0005). In view of the similarity, in all respects, of the patients who did and did not have IML, it is not surprising that the patients in whom IML was not carried out present a prognosis intermediate between those of patients with normal and with abnormal IML. It should be noted that the incidence of internal mammary involvement is completely disregarded in all prospective trials of adjuvant therapy. Results have sometimes been conflicting and disappointing. It is possible that, in many instances, unrecognised internal mammary lymphatic metastases have confounded the results of these studies. Arguments against routine internal mammary lymph node biopsy have often placed emphasis on the low incidence of internal mammary metastases without concomitant axillary involvement. The essential point at issue is the significantly reduced survival rate when the internal mammary nodes are involved in addition to the axilla. Veronesi and Valagussa (1983) have reported that involvement of both axillary and internal mammary nodes results in a 10-year survival of 20.5%, in contrast to 45.8% when only the internal mammary nodes were involved. Similar results have been reported by earlier investigators (Bucalossi et al., 1971; Haagenson, 1971;

471

Urban and Marjani, 1971; Livingston and Arlen, 1974; Handley, 1975; Veronesi and Valagussa, 1980. There has, therefore, been increasing awareness of the need to determine the status of the internal mammary lymphatics for a more comprehensive approach to therapeutic planning (Weichselbaum et al., 1976) and predicting the outcome of disease (Carter, 1979). Using radiocolloid, India ink or patent blue injection of the breast, several studies have shown incontrovertibly that no quadrant distinctions are warranted regarding lymphatic drainage to the axillary and internal mammary nodes (Hultborn et al., 1955; Turner-Warwick, 1958; Vendrell-Torn6 et al., 1972) and surgical studies have confirmed vulnerability of both sites to metastases from a tumour arising in any part of the breast (Bucalossi et al., 1971; Haagenson, 1971; Urban and Marjani, 1971; Livingston and Arlen, 1974; Handley, 1975; Veronesi and Valagussa, 1981). Veronesi et al. (1983) have proposed a method of calculating the risk of internal mammary lymph node metastases based on the age of the patient, the size of the primary tumour and the axillary node status. We propose the routine use of radiocolloid lymphoscintigraphy in all patients with breast carcinoma, not just those with medial lesions, regardless of the axillary nodal status. After a decade of experience in internal mammary lymphoscintigraphy it would seem that a TNM staging system which incorporates the results of IML will bring us a step closer to a more rational assessment of the patient with breast carcinoma. While the T and M classifications should remain as established, it is proposed that the N classification be subdivided into N A (axillary node status) and N I (internal mammary lymphatic status). Only when there is no demonstrable abnormality in either lymphatic area should the patient be considered to be at Stage I. Stage II would indicate involvement of one or other lymphatic area and Stage III, involvement of both axillary and internal mammary lymphatics. Lymphoscintigraphy is a simple, practicable, well tolerated method of examination, the results of which lie well within the limits of accuracy obtainable for many diagnostic procedures. Systematic inclusion of IML in prospective trials, in addition to the other currently recognised prognostic variables, would undoubtedly lead to a better understanding of the natural history of the disease and yield more definitive results by unravelling much of the conflicting evidence reported from prospective studies. Acknowledgement.We wish to acknowledge our surgical colleagues and those in the Departments of Radiation and Medical Oncology whose patients comprise the study group. The staff of the Department of Nuclear Medicine provided technical assistance. We are grateful to Mrs J. Reid and Mr P. Mlceli for the statistical analyses. Mrs O. Harasym-Malynowsky prepared the manuscript.

REFERENCES Bucalossi, P., Veronesl, U., Zmgo, L. & Cantu, C (1971). Enlarged mastectomy for breast cancer: review of 1,213 cases. American Journal of Roentgenology, 111, 119-122. Carter, S K. (1979). Importance of controlling the regional drainage lymph nodes in breast cancer. In Cancer Climcal Trtals, ed. Brady, L. W & DeVita, V., Vols 1, 2, pp. 227-233. Masson Publishing, New York. Ege, G. N. (1976). Internal mammary lymphoscintigraphy - the

472

CLINICAL RADIOLOGY

rationale, technique, interpretation and chnical application a rewew based on 848 cases. Radiology, 118, 101-107 Ege, G. N. (1983) Lymphoscintlgraphy - techniques and apphcatlons m the management of breast carcinoma Seminars tn Nuclear Medicine, 13, 2(3-34. Ege, G. N. & Clark, R M (1980). Internal mammary lymphoscmtlgraphy in the conservative surgical management of breast carcinoma. CIincal Radiology, 31,559-563. Ege, G N & Elhakim, T. (1984). The relevance of Internal mammary lymphoscintigraphy (1ML) in the management of breast carcinoma. Journal of Climcal Oncology, 2, 774-78l. Haagenson, C. D. (1971). Diseases of the Breast, 2rid edn, pp 413426. W. B. Saunders, Philadelphia. Handley, R. S (1975). Carcinoma of the breast Annals of the Royal College of Surgeons of England, 57, 59-66. Hill, N. S , Ege, G. N , Greyson, N. D , Mahoney, L. J & Jirsch, D. W (1983). Prediction by lymphoscintlgraphy ot nodal metastases in breast cancer. Canadian Journal of Surgery, 26,507 509. Hultborn, K. A., Larsson, L G . & Ragnhult, I. (1955). The lymph drainage from the breast to the axdlary and parasternal lymph nodes studied with the aid of colloidal Au-198. Acta Radiologlca, 43, 52-64. Livingston, S. F & Aden, M. (1974). The extended extrapleural radical mastectomy: its role in the treatment of carcinoma o1 the breast. Annals of Surgery, 179, 260-265 Matsuo, S. (/974). Studies of the metastasis of breast cancer to lymph

nodes - 11 Diagnosis ot metastasis to internal mammary nodes using radigcollold. Acta Medica Okayama, 28, 361-371. Shibata, H R. (1983). Editorial l~/mphoscintigraphy in the staging of breast cancer Canadian Journal of Surgery, 26, 487-488 Turner-Warwick, R. T (1958). The lymphatics of the breast. Brmsh Journal of Surgery, 46, 574-582. Urban, J A. & Marjani, M. A. (1971). Significance of internal mammary lymph node metastases in breast cancer. American Journal of RoentgenoIogy, 111, 130-136. Vendrell-Tornd, E., Setoain-Quinquer, J. & Domdnech-Torn~, F. M (1972). Study of normal mammary lymphatic drainage using radioactive isotopes. Journal of Nuclear Medicine, 13, 801-805, Veronesi, U , Cascinelll, N., Bufalino, R., Morablto, A., Greco, M., Galluzzo, D (1983) Risk of internal mammary lymph node metastases and its relevance on prognosis of breast cancer patients. Annals of Surgery, 198, 681-684. Veronesi, U & Valagussa, P. (1981) Inefficacy of internal mammary node dissection in breast cancer surgery Cancer, 47, 170-175. Veronesl, U & Valagussa, P. (1983). Internal mammary node involvement. Results of a prospective randomized study comparing two different surgical procedures. In Conservative Management

of Breast Cancer. New Surgical and Radtotherapeutzc Techmques, ed. Harris, J. R., Hellman, S. & Sden, W , pp. 201-209. J. B. LippIncott, Philadelphia. Weichselbaum, R. R , Marck, A. & Hellman , S (1976). The role of postoperative radiation in carcinoma of the breast. Cancer, 37, 2682-2690