- Email: [email protected]
Liver metastases from breast cancer: Long-term survival after curative resection
Liver metastases from breast cancer: Long-term survival after curative resection Markus Selzner, MD, Michael A. Morse, MD, James J. Vredenburgh, MD, William C. Meyers, MD, and Pierre-Alain Clavien, MD, PhD, Durham, NC
Background. Liver metastases from breast cancer are associated with a poor prognosis (median survival <6 months). A subgroup of these patients with no dissemination in other organs may benefit from surgery. Available data in the literature suggest that only in exceptional cases do these patients survive more than 2 years when given chemohormonal therapy or supportive care alone. We report the results of liver resection in patients with isolated hepatic metastases from breast cancer and evaluate the rate of long-term survival, prognostic factors, and the role of neoadjuvant high-dose chemotherapy. Patients and methods. Over the past decade, 17 women underwent hepatic metastectomy with curative intent for metastatic breast cancer. The follow-up was complete in each patient. The median age at the time breast cancer was diagnosed was 48 years. Neoadjuvant high-dose chemotherapy (HDC) with hematopoietic progenitor support was used in 10 patients before liver resection. Perioperative complications, long-term outcome, and prognostic factors were evaluated. Results. Seven of the 17 patients are currently alive, with follow-up of up to 12 years. Four of these patients are free of tumors after 6 and 17 months and 6 and 12 years. The actuarial 5-year survival rate is 22%. One patient died postoperatively (mortality rate, 6%) of carmustine-induced fibrosing pneumonitis. There was no further major morbidity in the other patients. The liver was the primary site of recurrent disease after liver resection in 67% of the patients. Patients in whom liver metastases were found more than 1 year after resection of the primary breast cancer had a significantly better outcome than those with early (<1 year) metastatic disease (P = .04). The type of liver resection, the lymph node status at the time of the primary breast cancer resection, and HDC had no significant impact on patient survival in this series. Conclusions. Favorable 22% long-term survival can be achieved with metastasectomy in this selected group of patients. Careful evaluation of pulmonary toxicity from carmustine and exclusion of patients with extrahepatic disease are critical. Improved survival might be achieved with better selection of patients and the use of liver-directed adjuvant therapy. (Surgery 2000;127:383-9.) From the Department of Surgery, Section of Hepatobiliary Surgery and Transplantation, and the Department of Medicine, Division of Hematology and Oncology, Duke University Medical Center, Durham, NC
LIVER METASTASES develop in approximately half of all women with metastatic breast cancer and are typically associated with tumor deposits at other sites, indicating advanced disease and poor outcome. However, in a small proportion of patients with breast cancer (1%), focal liver metastasis is the only sign of dissemination of the disease.1,2 The natural history of this condition is poorly defined, and the management remains controversial. Although most physicians view liver metastases Accepted for publication October 2, 1999. Reprint requests: Pierre-Alain Clavien, MD, PhD, FACS, Department of Surgery, Duke University Medical Center, PO Box 3247, Durham, NC 27710. Copyright © 2000 by Mosby, Inc. 0039-6060/2000/$12.00 + 0 doi:10.1067/msy.2000.103883
11/56/103883
from breast cancer with resignation or attempt palliation with hormones and chemotherapy, some groups have advocated aggressive approaches, including liver resection. The number of reported series has remained small and heterogeneous and, with only a few cases in each series, interpretation of the results is difficult.1-5 Liver resection offers the only chance of cure in patients with a variety of primary and secondary liver tumors; for example, 25% to 38% of patients with hepatic colorectal metastases are cured by surgery in the absence of extrahepatic disease.6-8 On this basis, a decade ago we opted for an aggressive policy, including hepatectomy, for women in whom liver metastases from breast cancer were the only manifestations of dissemination. Here we report the results in 17 consecutive women with resectable liver metastases from breast cancer seen SURGERY 383
384 Selzner et al
Fig 1. Portable anteroposterior chest radiograph shows heterogeneous reticulonodular opacities in both lungs. Multiple surgical clips are in the right axilla from the previous nodal resection and mastectomy.
Fig 2. Chest CT (lung window) reveals ground-glass opacities (arrows) and thickening of the interlobular septa with associated architectural distortion. These findings are consistent with acute alveolitis or fibrosis.
at Duke University Medical Center over the past decade. Ten of these patients received high-dose chemotherapy (HDC) and hematopoetic support before liver resection. The purpose of this study was to evaluate possible survival benefits, prognostic factors, and the role of HDC and hematopoietic support used as neoadjuvant therapy in this selected population of patients. PATIENTS AND METHODS Between August 1987 and January 1999, 6041 patients with breast cancer were treated at our institution. One percent of the patients presented with liver metastasis as the only manifestation of the disease. However, in only 33 patients was the liver believed to be resectable. Of these 33 patients, 17 underwent liver resection. The other 16 patients underwent exploratory procedures, but their livers were found to be unresectable at the time of surgery. The median age of these women at the
Surgery April 2000
time of breast surgery was 48 years (range, 32-72 years). Fifteen patients underwent a modified radical mastectomy, one patient had a simple mastectomy, and one had a lumpectomy. All patients underwent axillary lymph node dissection, and the number of lymph nodes obtained is known in each case. Histologic sections demonstrated infiltrating ductal adenocarcinoma in each case. Twelve of the 17 patients (71%) had positive axillary nodes (Table I). No patient had evidence of distant metastases at the time the primary breast cancer was diagnosed. Metastases subsequently developed in 3 of the 17 patients (lung 42 months, spine 24 months, and chest wall 37 months after lumbar breast surgery) before the liver metastases; these were treated with chemotherapy (all cases) and radiation therapy (chest wall and lumbar spine metastases), with complete response. All patients had preoperative CT or MRI evaluation of the chest and the abdomen. In addition, intraoperative ultrasonography was performed in each case, and since 1995 we have performed a positron-emission tomography (PET) scan in each patient. There was no evidence of extrahepatic disease at the time of liver resection in this series. The median time between resection of the primary breast tumor and diagnosis of liver metastases was 29 months (range, 2-84 months). Ten patients (59%) received induction chemotherapy with doxorubicin, 5-fluorouracil, and methotrexate, followed by high-dose chemotherapy of cyclophosphamide, cisplatin, and carmustine (BCNU) combined with hematopoietic progenitor support before liver resection (Table I). The median time between HDC and liver resection was 8 months (range, 3-24 months). The decision to proceed with HDC in these patients was based on the response to induction chemotherapy (doxorubicin, 5-fluorouracil, and methotrexate). Partial response to HDC was defined as a decrease in tumor size of about 50%. The three patients with previous extrahepatic disease had no evidence of disease outside the liver after HDC. Although all patients experienced moderate side effects from chemotherapy, including nausea and vomiting, no life-threatening complications were noted in this series. Three patients in whom estrogen receptors were present in the metastatic tissue received adjuvant antiestrogen therapy after liver resection. Of the 17 patients, 12 (71%) had a single liver metastasis and 5 (29%) had two lesions. The median size of the metastases was 2.5 cm (range, 1.5-5 cm). All tumors were believed to be resectable with a curative intent (ie, 1-cm negative tumor margin).
Surgery Volume 127, Number 4
Selzner et al 385
Fig 3. Actuarial survival after curative liver resection for metastatic breast cancer (n = 17). Twenty-two percent of the patients are alive after 5 years, with 17% being free of disease.
Fig 4. Survival of patients with liver recurrence within 1 year (broken line, n = 4) and after 1 year (solid line, n = 13) of breast surgery. Patients with early liver metastases had a significanly shorter survival than patients whose cancer recurred after 1 year (P = .04, “log rank” test).
Survival analysis was performed with the “log rank” test. Significance was declared when P was <.05.
overall median survival of the 17 patients was 24 months after liver resection, with 7 patients currently alive at 6, 12, 14, 17 months and 2, 6, and 12 years. Four of them have no evidence of recurrent disease at 6 and 17 months and at 6 and 12 years. The median survival time of the 16 patients with curative resection was 27 months, with an actuarial 5-year survival rate of 22% and a 5-year disease-free survival rate of 17% (Fig 3). In the 12 patients in whom recurrent disease developed, the median time between liver resection and the first sign of recurrence was 7 months (range, 1 to 24 months). The liver was the first site of recurrent disease in eight patients (67%), whereas in four other patients (36%) bone was the first site of recurrence. Patients given HDC with hematopoietic support before liver resection (n = 10) showed no apparent benefit in terms of overall or disease-free survival, when compared with the seven patients who underwent surgery alone. Of note, the two patients with long-term survival and no evidence of disease after 6 and 12 years of follow-up did not receive HDC before liver resection. Several prognostic factors were investigated in this series. The median survival was only 9 months in four patients in whom liver metastases developed within 1 year of breast surgery compared with 27 months in those with metastatic disease diagnosed more than 12 months after the initial diagnosis (P = .04, log rank test) (Fig 4). Excluding the patient with a positive lymph node in the porta hepatis, the period between breast surgery and liver metastases still shows a trend but is no longer
RESULTS Hepatic metastases were resected by anatomic (n = 7) or wedge resection (n = 10), with negative margins in all cases (Table I). The absence of additional hepatic involvement was routinely documented by intraoperative ultrasonography. In one woman a positive lymph node was documented in the porta hepatis, and the liver resection was considered noncurative. This woman had early recurrence within the liver and died 5 months postoperatively. The median operative blood loss was 350 mL (150-2000 mL), with only four patients requiring blood transfusions (all during the first 5 years of the study). In one patient a fulminate ARDS syndrome developed 5 days after an uneventful right hepatectomy. Despite rapid tracheal reintubation, maximal therapy in the intensive care unit, and the administration 2 g IV Solu-Medrol (methylprednisolone), she died within 24 hours of respiratory insufficiency and right-side heart failure (pulmonary arterial pressure 90/60, partial arterial oxygen pressure <60 mm Hg on 100% oxygen and a positive end-expiratory pressure [PEEP] of 10 mm Hg, and a cardiac output <2 L). An autopsy demonstrated fibrosing pneumonitis, most likely induced by BCNU chemotherapy (Figs 1 and 2). There was no major morbidity in the other patients, and the median hospital stay was 7 days (range, 4-10 days). Each patient had a complete follow-up, with a median of 17 months (6 months to 12 years). The
386 Selzner et al
Surgery April 2000
Table I. Characteristics of the patients treated with liver resection for metastatic breast cancer Patient age Patient (y) 1 2 3 4 5 6 7 8* 9 10 11 12 13 14 15 16 17
62 43 48 40 34 57 51 50 40 32 52 35 69 72 41 66 48
Tumor size of BC (cm) 2.5 2.3 3.5 1.5 2.2 1.3 2.2 2.2 6 3 2 3 4 2.5 2.5 2.3 1.5
LN-status in the axilla at breast surgery (pos/total) 0/8 5/10 14/28 3/13 1/17 5/8 0/12 0/16 0/13 10/24 2/20 2/15 2/9 0/15 1/22 1/7 1/19
Time between breast surgery and diagnosis of liver metastases (mo) 35 2 70 58 21 40 13 9 84 23 9 39 11 24 19 12 41
Extrahepatic disease No No No No No No No Lung Chest wall No No Lumbar spine No No No No No
HDC Yes Yes No Yes Yes No No Yes Yes No Yes Yes No No Yes No Yes
HDC, High-dose chemotherapy with hematopoietic support; WR, wedge resection; HH, hemihepatectomy. *Positive lymph node in the porta hepatis at the time of liver resection. †Death due to carmustine-related fulminate fibrosing pneumonitis 5 days after an uneventful right hepatectomy.
statistically significant (P = .1). The three patients with extrahepatic metastatic disease before liver resection showed no statistically significant decreased survival (5, 24, and 57 months) compared with the patients who did not have extrahepatic disease before chemotherapy. No significant differences were found in overall and recurrence-free survival between patients treated with wedge and segmental resection (n = 10; median recurrence-free survival, 5 months) versus anatomic hemihepatectomy (n = 6; median survival, 6 months). However, patients who underwent anatomic resection had larger metastases than patients treated with wedge or segmental resection (4 cm vs 2.5 cm; P = .02; Student t-test). Patients with negative axillary lymph nodes at the time of breast surgery (n = 5) had no statistically significant difference in the length of disease-free survival compared with those with positive lymph nodes (n = 12). However, both long-term survivors (6 and 12 years) had no involvement of the axillary lymph nodes at the time of breast surgery. The size or number of metastases in the liver (1 or 2) and the presence of estrogen or progesterone receptors had no significant impact on survival in this series. DISCUSSION Breast cancer associated with visceral metastases is usually considered a systemic disease with a poor prognosis. Isolated liver metastasis as the only manifestation of recurrence, although rare, might be a
more favorable condition suitable for an aggressive protocol that includes liver resection. Because the mortality and morbidity associated with liver resection have decreased dramatically during the last decade,6-8 the indication for surgery in a variety of metastatic diseases has widened.9 However, the effectiveness of liver resection for noncolorectal and non-neuroendocrine metastases remains controversial. This study suggests that metastasectomy might be of significant benefit in a select group of women with breast cancer metastatic to the liver. Isolated liver metastases from breast cancer are rare, and the number of series reported in the literature is small and heterogeneous. The natural history of this condition and the value of various therapies remain elusive. From the few reported series, the median survival in this select group of patients treated with hormones and chemotherapy or supportive care ranges between 3 and 6 months, with only exceptional documented survival for more than 2 years after the diagnosis.10-15 This apparently is true even in patients with isolated small single metastases to the liver. This population of patients most closely resembles the patients who underwent liver resection in our series. Therefore these data may serve as the basis for comparison of the results of other therapeutic modalities, including surgery. We report here the largest series of liver resections in women with metastatic breast cancer with a complete follow-up (Table II). All patients were
Selzner et al 387
Surgery Volume 127, Number 4
Type of liver resection
Size of liver metastases (cm)
Number of liver metastases
Time between liver resection and recurrent disease (mo)
Death after liver resection (mo)
Alive follow-up
2.3 2 1.9 2 1.5 2 2.5 4.4 2.5 2.5 2.3 5 4 4 4 3 2.7
1 2 2 2 1 1 1 1 1 1 1 1 1 1 2 1 2
5 5 6 1 6 11 No recurrence 3 12 6 3 23 7 No recurrence No recurrence No recurrence —
27 5 — 14 27 23 — 5 — 15 15 57 — — — — Post OP†
— — 14 mo — — — 6 yr — 24 mo — — — 12 mo 12 yr 6 mo 17 mo —
WR WR WR HH WR WR WR HH HH WR WR Ext. HH HH HH WR WR HH
Table II. Reported series of patients treated with liver resection for metastatic breast cancer with or without adjuvant chemotherapy Authors/Year Foster et al,3 1978 Stehlin et al,17 1988 Szakacs et al,18 1990 Elias et al,4 1991 Wolf et al,19 1991 Schneebaum et al,15 1994 Lorenz et al,1 1995 Raab et al,16 1998 Duke, 1999
No. of patients 5 9 1 12 1 6 8 30 17
Treatment
Death (mo)
Resection Resection + IV chemotherapy Resection Resection + IV chemotherapy Resection Resection + art. chemotherapy Resection + art. chemotherapy Resection Resection + IV chemotherapy
<1/3/6/8/19 5 9
Median survival (mo) 5-Year survival NA 28 NA 37 NA 42 15 41 27
NA NA NA NA NA NA NA 22% 22%
NA, Not available.
treated and followed up in a single institution. Several patients had long-term survival, with disease-free survival documented at 6 and 12 years after surgery. The overall 22% actuarial survival rate at 5 years compares favorably with the outcome of other approaches, including hormonechemotherapy or supportive care only. This survival is also surprisingly in the range of reported series of patients with colorectal cancer metastatic to the liver. However, although the overall results may appear encouraging, in most patients (75%) recurrence develops in a median of 8 months after surgery. These data provide a strong impetus to identify prognostic factors for selecting the group of patients who are most likely to benefit from
surgery. The high rate of early hepatic recurrence after hepatectomy should also focus our attention on innovative adjuvant therapy protocols. A few other studies have suggested that patients with isolated liver metastases from breast cancer may benefit from liver resection (Table II). Raab et al16 reported a similar survival rate in a series of 30 patients treated with curative resection. However, incomplete data were provided regarding recurrent disease, and no long-term disease-free survival was reported since the follow-up was performed only by mail questionnaire. The authors also did not evaluate prognostic factors, and no information was provided on adjuvant chemotherapy. Elias et al4 reported a median survival of 37 months in
388 Selzner et al
12 patients with breast cancer metastatic to the liver who were treated with metastectomy. Also, no data were provided on disease-free survival, and no prognostic factors were identified. In addition, it was unclear in this series whether each patient was treated with curative resection. Stehlin et al17 compared nine patients whose metastatic breast cancer was treated with liver resection and chemotherapy with a historical series of 62 patients treated with chemotherapy alone. Patients treated with resection plus chemotherapy had a significantly prolonged survival compared with patients who received chemotherapy alone (28 months vs 5 months). In contrast, several other small series of fewer than five patients3,18,19 failed to show any survival advantage in patients treated with hepatic resection for metastatic breast cancer. In this series, we described a patient in whom a fulminant and lethal interstitial pneumonitis developed 5 days after curative resection of a liver metastasis. This type of injury is a well-described complication of BCNU therapy. The incidence of interstitial pneumonitis (ARDS syndrome) in patients receiving BCNU as part of the HDC regimen has been reported to be between 39% and 53%.20,21 Previous radiation therapy of the chest,22 pre-existing lung disease,23 and a history of smoking24 have been identified as additional risk factors. The pathogenesis of this respiratory failure has not been extensively studied, but evidence has been provided that pulmonary macrophages can produce excessive amounts of collagen in response to high partial pressure of oxygen (PO2),23 resulting in rapidly developing interstitial fibrosis and ARDS syndrome.22,25 This condition is usually reversible with rapid IV administration of steroid (eg, 60 mg prednisone twice a day until resolution of the pulmonary symptoms). In our patient, large doses of steroid (Solu-Medrol 2 g) did not reverse the fulminant course of the respiratory failure and had no effect on pulmonary hypertension and the rightside heart failure. This patient is the last patient of our series, and now we routinely use prophylactic steroid (2 × 60 mg prednisone) in each patient previously exposed to BCNU, and we restrict the inspiratory oxygen concentration to less than 30% during and after surgery. In patients with diffusion capacity <60% in preoperative pulmonary function tests, we also give postoperative steroids for 14 days (20 mg every 12 hours). The high incidence of recurrence within the liver after curative resection (67%) indicates either that biologic factors, yet unknown, might predispose the liver as a favorable site of seeding in selected types of breast cancer or that micrometastases
Surgery April 2000
had remained undetected at the time of resection (67%). New imaging modalities such as the PET scan might be useful to enhance detection of small metastases throughout the liver.26 The use of selective adjuvant chemotherapy delivered to the liver with a subcutaneous pump device via a catheter inserted into the gastroduodenal artery may reduce the risk of early intrahepatic recurrence. Several prognostic factors were investigated in this series of 17 patients. The only significant factor identified was the time between primary breast surgery and the diagnosis of liver metastases. Patients in whom liver metastases developed within the first year after primary breast surgery had a significantly shorter survival (median survival, 9 months vs 27 months; P = .04 log rank test). Although aggressive preoperative treatment with ablative HDC was not associated with a statistically significant improvement in patient survival when compared with standard chemotherapy regimens, a benefit from HDC cannot be excluded. A bias in the selection of patients for preoperative therapy is unavoidable in this type of study. Similarly, although the lymph node status at the time of breast surgery had no significant influence on patient survival in this series, the role of initial staging of the breast cancer may be important. Both of the long-term survivors in our series had no lymph node involvement in the axilla. In summary, liver resection can be performed with minimal morbidity and low mortality in this patient group. Pulmonary toxicity from previous chemotherapy using BCNU should be carefully evaluated before surgery, and prophylactic steroids should be administered along with minimizing PO2. Liver resection in this selected group is associated with a favorable 22% 5-year survival rate. Patients with a long interval (more than 1 year) between breast cancer resection and liver metastases are most likely to benefit from surgery. Additional adjuvant therapies, such as selective adjuvant intra-arterial chemotherapy, should be investigated to decrease the high incidence of hepatic recurrence after resection. REFERENCES 1. Lorenz M, Wiesner J, Staib-Sebler E, Encke A. Regionale Therapie von Mammakarzinom-Lebermetastasen. Zentralbl Chir 1995;120:786-90. 2. Raab R, Nussbaum KT, Pichlmayr R. Lebermetastasen bei Mammacarcinom. Chirurg 1996;67:234-7. 3. Foster JH. Survival after liver resection for secondary tumors. Am J Surg 1978;135:389-93. 4. Elias D, Lasser P, Spielmann M, et al. Surgical and chemotherapeutic treatment of hepatic metastases from carcinoma of the breast. Surgery 1991;172:461-4.
Surgery Volume 127, Number 4 5. Pocard M, Salmon R-J. Resection hepatique pour metastase du cancer du sein: le concept de chirurgie adjuvante. Bull Cancer 1997;84:47-50. 6. Selzner M, Clavien P-A. Resection of liver tumors: special emphasis on neoadjuvant and adjuvant therapy. In Clavien P-A, ed. Malignant Liver Tumors. Malden, Mass: Blackwell Science; 1999:137-149. 7. Fong Y, Cohen A, Fortner J, et al. Liver resection for colorectal metastases. J Clin Oncol 1997;15:938-46. 8. Gayowski TJ, Iwatsuki S, Madariaga JR. Experience in hepatic resection for metastatic colorectal cancer: analysis of clinical and pathologic risk factors. Surgery 1994;116:703. 9. Selzner M, Clavien P-A. Liver tumors of rare and unknown origins. In Clavien P-A, ed. Malignant Liver Tumors. Malden, Mass: Blackwell Science; 1999:303-312. 10. Hoe A, Royle G, Taylor I. Breast cancer and liver metastases— incidence, diagnosis and outcome. J R Soc Med 1984;84:714-6. 11. Zinser J, Hortobagyi G, Buzdar A, et al. Clinical course of breast cancer patients with liver metastases. J Clin Oncol 1987;5:773-82. 12. Patanaphan V, Salazar M, Risco R. Breast cancer: metastatic patterns and their prognosis. South Med J 1988;81:1109-12. 13. O’Reilly S, Richards M, Rubens R. Liver metastases from breast cancer: the relationship between clinical, biochemical and pathological features and survival. Eur J Cancer 1990;26:574-7. 14. Zimmermann T, Padberg W, Kelm C, et al. Die lokoregionale Chemotherapie zur Behandlung von Lebermetastasen des Mammakarzinoms. Zentralbl Chir 1992;117:226-30. 15. Schneebaum S, Walker MJ, Young D, et al. The regional treatment of liver metastases from breast cancer. J Surg Oncol 1994;55:26-32. 16. Raab R, Nussbaum KT, Behrend M, Weimann A. Liver metastases of breast cancer—results of liver resection. Anticancer Res 1998;18(3C):2231-3.
Selzner et al 389
17. Stehlin JS, Ipolyi PD, Greeff PJ, et al. Treatment of cancer of the liver. Ann Surg 1988;208:23-35. 18. Szakacs JG, Szakacs JE, Karl RC. Surgical resection versus perfusion in the treatment of metastatic and primary liver tumors. Ann Clin Lab Sci 1990;20:245-56. 19. Wolf RF, Goodnight JE, Krag DE, Schneider PD. Results of resection and proposed guidelines for patient selection in instances of noncolorectal hepatic metastases. Surg Gynecol Obstet 1991;173:454-60. 20. Chap L, Shipner R, Levine M, et al. Pulmonary toxicity of high-dose chemotherapy for breast cancer: a non-invasive aproach to diagnosis and treatment. Bone Marrow Transplant 1997;20:1063-7. 21. Wilczynski S, Erasmus J, Petros W, et al. Delayed pulmonary toxicity syndrome following high-dose chemotherapy and bone marrow transplantation for breast cancer. Am J Resp Crit Care Med 1998;157:565-73. 22. Todd N, Peters W, Ost A, et al. Pulmonary drug toxicity in patients with primary breast cancer treated with high-dose combination chemotherapy and autologous bone marrow transplantation. Am Rev Respir Dis 1993;147:1264-70. 23. Cherniack R, Abrams J, Kalica A. Pulmonary disease associated with breast cancer therapy. Am J Respir Crit Care Med 1994;150:1169-73. 24. Demirer T, Weaver C, Buckner C, et al. High-dose cyclophosphamide, carmustine and etoposide followed by allogeneic bone marrow transplantation in ptients with lymphoid malignancies who received prior dose-limiting radiation therapy. J Clin Oncol 1995;13:596-602. 25. Smith A. The pulmonary toxicity of nitrosoureas. Pharmacol Ther 1989;41:443-60. 26. Bender H, Kirst J, Palmedo H, et al. Value of 18 fluorodeoxyglucose positron emission tomography in the staging of recurrent breast carcinoma. Anticancer Res 1997; 17(3B):1687-92.
Background. Liver metastases from breast cancer are associated with a poor prognosis (median survival <6 months). A subgroup of these patients with no dissemination in other organs may benefit from surgery. Available data in the literature suggest that only in exceptional cases do these patients survive more than 2 years when given chemohormonal therapy or supportive care alone. We report the results of liver resection in patients with isolated hepatic metastases from breast cancer and evaluate the rate of long-term survival, prognostic factors, and the role of neoadjuvant high-dose chemotherapy. Patients and methods. Over the past decade, 17 women underwent hepatic metastectomy with curative intent for metastatic breast cancer. The follow-up was complete in each patient. The median age at the time breast cancer was diagnosed was 48 years. Neoadjuvant high-dose chemotherapy (HDC) with hematopoietic progenitor support was used in 10 patients before liver resection. Perioperative complications, long-term outcome, and prognostic factors were evaluated. Results. Seven of the 17 patients are currently alive, with follow-up of up to 12 years. Four of these patients are free of tumors after 6 and 17 months and 6 and 12 years. The actuarial 5-year survival rate is 22%. One patient died postoperatively (mortality rate, 6%) of carmustine-induced fibrosing pneumonitis. There was no further major morbidity in the other patients. The liver was the primary site of recurrent disease after liver resection in 67% of the patients. Patients in whom liver metastases were found more than 1 year after resection of the primary breast cancer had a significantly better outcome than those with early (<1 year) metastatic disease (P = .04). The type of liver resection, the lymph node status at the time of the primary breast cancer resection, and HDC had no significant impact on patient survival in this series. Conclusions. Favorable 22% long-term survival can be achieved with metastasectomy in this selected group of patients. Careful evaluation of pulmonary toxicity from carmustine and exclusion of patients with extrahepatic disease are critical. Improved survival might be achieved with better selection of patients and the use of liver-directed adjuvant therapy. (Surgery 2000;127:383-9.) From the Department of Surgery, Section of Hepatobiliary Surgery and Transplantation, and the Department of Medicine, Division of Hematology and Oncology, Duke University Medical Center, Durham, NC
LIVER METASTASES develop in approximately half of all women with metastatic breast cancer and are typically associated with tumor deposits at other sites, indicating advanced disease and poor outcome. However, in a small proportion of patients with breast cancer (1%), focal liver metastasis is the only sign of dissemination of the disease.1,2 The natural history of this condition is poorly defined, and the management remains controversial. Although most physicians view liver metastases Accepted for publication October 2, 1999. Reprint requests: Pierre-Alain Clavien, MD, PhD, FACS, Department of Surgery, Duke University Medical Center, PO Box 3247, Durham, NC 27710. Copyright © 2000 by Mosby, Inc. 0039-6060/2000/$12.00 + 0 doi:10.1067/msy.2000.103883
11/56/103883
from breast cancer with resignation or attempt palliation with hormones and chemotherapy, some groups have advocated aggressive approaches, including liver resection. The number of reported series has remained small and heterogeneous and, with only a few cases in each series, interpretation of the results is difficult.1-5 Liver resection offers the only chance of cure in patients with a variety of primary and secondary liver tumors; for example, 25% to 38% of patients with hepatic colorectal metastases are cured by surgery in the absence of extrahepatic disease.6-8 On this basis, a decade ago we opted for an aggressive policy, including hepatectomy, for women in whom liver metastases from breast cancer were the only manifestations of dissemination. Here we report the results in 17 consecutive women with resectable liver metastases from breast cancer seen SURGERY 383
384 Selzner et al
Fig 1. Portable anteroposterior chest radiograph shows heterogeneous reticulonodular opacities in both lungs. Multiple surgical clips are in the right axilla from the previous nodal resection and mastectomy.
Fig 2. Chest CT (lung window) reveals ground-glass opacities (arrows) and thickening of the interlobular septa with associated architectural distortion. These findings are consistent with acute alveolitis or fibrosis.
at Duke University Medical Center over the past decade. Ten of these patients received high-dose chemotherapy (HDC) and hematopoetic support before liver resection. The purpose of this study was to evaluate possible survival benefits, prognostic factors, and the role of HDC and hematopoietic support used as neoadjuvant therapy in this selected population of patients. PATIENTS AND METHODS Between August 1987 and January 1999, 6041 patients with breast cancer were treated at our institution. One percent of the patients presented with liver metastasis as the only manifestation of the disease. However, in only 33 patients was the liver believed to be resectable. Of these 33 patients, 17 underwent liver resection. The other 16 patients underwent exploratory procedures, but their livers were found to be unresectable at the time of surgery. The median age of these women at the
Surgery April 2000
time of breast surgery was 48 years (range, 32-72 years). Fifteen patients underwent a modified radical mastectomy, one patient had a simple mastectomy, and one had a lumpectomy. All patients underwent axillary lymph node dissection, and the number of lymph nodes obtained is known in each case. Histologic sections demonstrated infiltrating ductal adenocarcinoma in each case. Twelve of the 17 patients (71%) had positive axillary nodes (Table I). No patient had evidence of distant metastases at the time the primary breast cancer was diagnosed. Metastases subsequently developed in 3 of the 17 patients (lung 42 months, spine 24 months, and chest wall 37 months after lumbar breast surgery) before the liver metastases; these were treated with chemotherapy (all cases) and radiation therapy (chest wall and lumbar spine metastases), with complete response. All patients had preoperative CT or MRI evaluation of the chest and the abdomen. In addition, intraoperative ultrasonography was performed in each case, and since 1995 we have performed a positron-emission tomography (PET) scan in each patient. There was no evidence of extrahepatic disease at the time of liver resection in this series. The median time between resection of the primary breast tumor and diagnosis of liver metastases was 29 months (range, 2-84 months). Ten patients (59%) received induction chemotherapy with doxorubicin, 5-fluorouracil, and methotrexate, followed by high-dose chemotherapy of cyclophosphamide, cisplatin, and carmustine (BCNU) combined with hematopoietic progenitor support before liver resection (Table I). The median time between HDC and liver resection was 8 months (range, 3-24 months). The decision to proceed with HDC in these patients was based on the response to induction chemotherapy (doxorubicin, 5-fluorouracil, and methotrexate). Partial response to HDC was defined as a decrease in tumor size of about 50%. The three patients with previous extrahepatic disease had no evidence of disease outside the liver after HDC. Although all patients experienced moderate side effects from chemotherapy, including nausea and vomiting, no life-threatening complications were noted in this series. Three patients in whom estrogen receptors were present in the metastatic tissue received adjuvant antiestrogen therapy after liver resection. Of the 17 patients, 12 (71%) had a single liver metastasis and 5 (29%) had two lesions. The median size of the metastases was 2.5 cm (range, 1.5-5 cm). All tumors were believed to be resectable with a curative intent (ie, 1-cm negative tumor margin).
Surgery Volume 127, Number 4
Selzner et al 385
Fig 3. Actuarial survival after curative liver resection for metastatic breast cancer (n = 17). Twenty-two percent of the patients are alive after 5 years, with 17% being free of disease.
Fig 4. Survival of patients with liver recurrence within 1 year (broken line, n = 4) and after 1 year (solid line, n = 13) of breast surgery. Patients with early liver metastases had a significanly shorter survival than patients whose cancer recurred after 1 year (P = .04, “log rank” test).
Survival analysis was performed with the “log rank” test. Significance was declared when P was <.05.
overall median survival of the 17 patients was 24 months after liver resection, with 7 patients currently alive at 6, 12, 14, 17 months and 2, 6, and 12 years. Four of them have no evidence of recurrent disease at 6 and 17 months and at 6 and 12 years. The median survival time of the 16 patients with curative resection was 27 months, with an actuarial 5-year survival rate of 22% and a 5-year disease-free survival rate of 17% (Fig 3). In the 12 patients in whom recurrent disease developed, the median time between liver resection and the first sign of recurrence was 7 months (range, 1 to 24 months). The liver was the first site of recurrent disease in eight patients (67%), whereas in four other patients (36%) bone was the first site of recurrence. Patients given HDC with hematopoietic support before liver resection (n = 10) showed no apparent benefit in terms of overall or disease-free survival, when compared with the seven patients who underwent surgery alone. Of note, the two patients with long-term survival and no evidence of disease after 6 and 12 years of follow-up did not receive HDC before liver resection. Several prognostic factors were investigated in this series. The median survival was only 9 months in four patients in whom liver metastases developed within 1 year of breast surgery compared with 27 months in those with metastatic disease diagnosed more than 12 months after the initial diagnosis (P = .04, log rank test) (Fig 4). Excluding the patient with a positive lymph node in the porta hepatis, the period between breast surgery and liver metastases still shows a trend but is no longer
RESULTS Hepatic metastases were resected by anatomic (n = 7) or wedge resection (n = 10), with negative margins in all cases (Table I). The absence of additional hepatic involvement was routinely documented by intraoperative ultrasonography. In one woman a positive lymph node was documented in the porta hepatis, and the liver resection was considered noncurative. This woman had early recurrence within the liver and died 5 months postoperatively. The median operative blood loss was 350 mL (150-2000 mL), with only four patients requiring blood transfusions (all during the first 5 years of the study). In one patient a fulminate ARDS syndrome developed 5 days after an uneventful right hepatectomy. Despite rapid tracheal reintubation, maximal therapy in the intensive care unit, and the administration 2 g IV Solu-Medrol (methylprednisolone), she died within 24 hours of respiratory insufficiency and right-side heart failure (pulmonary arterial pressure 90/60, partial arterial oxygen pressure <60 mm Hg on 100% oxygen and a positive end-expiratory pressure [PEEP] of 10 mm Hg, and a cardiac output <2 L). An autopsy demonstrated fibrosing pneumonitis, most likely induced by BCNU chemotherapy (Figs 1 and 2). There was no major morbidity in the other patients, and the median hospital stay was 7 days (range, 4-10 days). Each patient had a complete follow-up, with a median of 17 months (6 months to 12 years). The
386 Selzner et al
Surgery April 2000
Table I. Characteristics of the patients treated with liver resection for metastatic breast cancer Patient age Patient (y) 1 2 3 4 5 6 7 8* 9 10 11 12 13 14 15 16 17
62 43 48 40 34 57 51 50 40 32 52 35 69 72 41 66 48
Tumor size of BC (cm) 2.5 2.3 3.5 1.5 2.2 1.3 2.2 2.2 6 3 2 3 4 2.5 2.5 2.3 1.5
LN-status in the axilla at breast surgery (pos/total) 0/8 5/10 14/28 3/13 1/17 5/8 0/12 0/16 0/13 10/24 2/20 2/15 2/9 0/15 1/22 1/7 1/19
Time between breast surgery and diagnosis of liver metastases (mo) 35 2 70 58 21 40 13 9 84 23 9 39 11 24 19 12 41
Extrahepatic disease No No No No No No No Lung Chest wall No No Lumbar spine No No No No No
HDC Yes Yes No Yes Yes No No Yes Yes No Yes Yes No No Yes No Yes
HDC, High-dose chemotherapy with hematopoietic support; WR, wedge resection; HH, hemihepatectomy. *Positive lymph node in the porta hepatis at the time of liver resection. †Death due to carmustine-related fulminate fibrosing pneumonitis 5 days after an uneventful right hepatectomy.
statistically significant (P = .1). The three patients with extrahepatic metastatic disease before liver resection showed no statistically significant decreased survival (5, 24, and 57 months) compared with the patients who did not have extrahepatic disease before chemotherapy. No significant differences were found in overall and recurrence-free survival between patients treated with wedge and segmental resection (n = 10; median recurrence-free survival, 5 months) versus anatomic hemihepatectomy (n = 6; median survival, 6 months). However, patients who underwent anatomic resection had larger metastases than patients treated with wedge or segmental resection (4 cm vs 2.5 cm; P = .02; Student t-test). Patients with negative axillary lymph nodes at the time of breast surgery (n = 5) had no statistically significant difference in the length of disease-free survival compared with those with positive lymph nodes (n = 12). However, both long-term survivors (6 and 12 years) had no involvement of the axillary lymph nodes at the time of breast surgery. The size or number of metastases in the liver (1 or 2) and the presence of estrogen or progesterone receptors had no significant impact on survival in this series. DISCUSSION Breast cancer associated with visceral metastases is usually considered a systemic disease with a poor prognosis. Isolated liver metastasis as the only manifestation of recurrence, although rare, might be a
more favorable condition suitable for an aggressive protocol that includes liver resection. Because the mortality and morbidity associated with liver resection have decreased dramatically during the last decade,6-8 the indication for surgery in a variety of metastatic diseases has widened.9 However, the effectiveness of liver resection for noncolorectal and non-neuroendocrine metastases remains controversial. This study suggests that metastasectomy might be of significant benefit in a select group of women with breast cancer metastatic to the liver. Isolated liver metastases from breast cancer are rare, and the number of series reported in the literature is small and heterogeneous. The natural history of this condition and the value of various therapies remain elusive. From the few reported series, the median survival in this select group of patients treated with hormones and chemotherapy or supportive care ranges between 3 and 6 months, with only exceptional documented survival for more than 2 years after the diagnosis.10-15 This apparently is true even in patients with isolated small single metastases to the liver. This population of patients most closely resembles the patients who underwent liver resection in our series. Therefore these data may serve as the basis for comparison of the results of other therapeutic modalities, including surgery. We report here the largest series of liver resections in women with metastatic breast cancer with a complete follow-up (Table II). All patients were
Selzner et al 387
Surgery Volume 127, Number 4
Type of liver resection
Size of liver metastases (cm)
Number of liver metastases
Time between liver resection and recurrent disease (mo)
Death after liver resection (mo)
Alive follow-up
2.3 2 1.9 2 1.5 2 2.5 4.4 2.5 2.5 2.3 5 4 4 4 3 2.7
1 2 2 2 1 1 1 1 1 1 1 1 1 1 2 1 2
5 5 6 1 6 11 No recurrence 3 12 6 3 23 7 No recurrence No recurrence No recurrence —
27 5 — 14 27 23 — 5 — 15 15 57 — — — — Post OP†
— — 14 mo — — — 6 yr — 24 mo — — — 12 mo 12 yr 6 mo 17 mo —
WR WR WR HH WR WR WR HH HH WR WR Ext. HH HH HH WR WR HH
Table II. Reported series of patients treated with liver resection for metastatic breast cancer with or without adjuvant chemotherapy Authors/Year Foster et al,3 1978 Stehlin et al,17 1988 Szakacs et al,18 1990 Elias et al,4 1991 Wolf et al,19 1991 Schneebaum et al,15 1994 Lorenz et al,1 1995 Raab et al,16 1998 Duke, 1999
No. of patients 5 9 1 12 1 6 8 30 17
Treatment
Death (mo)
Resection Resection + IV chemotherapy Resection Resection + IV chemotherapy Resection Resection + art. chemotherapy Resection + art. chemotherapy Resection Resection + IV chemotherapy
<1/3/6/8/19 5 9
Median survival (mo) 5-Year survival NA 28 NA 37 NA 42 15 41 27
NA NA NA NA NA NA NA 22% 22%
NA, Not available.
treated and followed up in a single institution. Several patients had long-term survival, with disease-free survival documented at 6 and 12 years after surgery. The overall 22% actuarial survival rate at 5 years compares favorably with the outcome of other approaches, including hormonechemotherapy or supportive care only. This survival is also surprisingly in the range of reported series of patients with colorectal cancer metastatic to the liver. However, although the overall results may appear encouraging, in most patients (75%) recurrence develops in a median of 8 months after surgery. These data provide a strong impetus to identify prognostic factors for selecting the group of patients who are most likely to benefit from
surgery. The high rate of early hepatic recurrence after hepatectomy should also focus our attention on innovative adjuvant therapy protocols. A few other studies have suggested that patients with isolated liver metastases from breast cancer may benefit from liver resection (Table II). Raab et al16 reported a similar survival rate in a series of 30 patients treated with curative resection. However, incomplete data were provided regarding recurrent disease, and no long-term disease-free survival was reported since the follow-up was performed only by mail questionnaire. The authors also did not evaluate prognostic factors, and no information was provided on adjuvant chemotherapy. Elias et al4 reported a median survival of 37 months in
388 Selzner et al
12 patients with breast cancer metastatic to the liver who were treated with metastectomy. Also, no data were provided on disease-free survival, and no prognostic factors were identified. In addition, it was unclear in this series whether each patient was treated with curative resection. Stehlin et al17 compared nine patients whose metastatic breast cancer was treated with liver resection and chemotherapy with a historical series of 62 patients treated with chemotherapy alone. Patients treated with resection plus chemotherapy had a significantly prolonged survival compared with patients who received chemotherapy alone (28 months vs 5 months). In contrast, several other small series of fewer than five patients3,18,19 failed to show any survival advantage in patients treated with hepatic resection for metastatic breast cancer. In this series, we described a patient in whom a fulminant and lethal interstitial pneumonitis developed 5 days after curative resection of a liver metastasis. This type of injury is a well-described complication of BCNU therapy. The incidence of interstitial pneumonitis (ARDS syndrome) in patients receiving BCNU as part of the HDC regimen has been reported to be between 39% and 53%.20,21 Previous radiation therapy of the chest,22 pre-existing lung disease,23 and a history of smoking24 have been identified as additional risk factors. The pathogenesis of this respiratory failure has not been extensively studied, but evidence has been provided that pulmonary macrophages can produce excessive amounts of collagen in response to high partial pressure of oxygen (PO2),23 resulting in rapidly developing interstitial fibrosis and ARDS syndrome.22,25 This condition is usually reversible with rapid IV administration of steroid (eg, 60 mg prednisone twice a day until resolution of the pulmonary symptoms). In our patient, large doses of steroid (Solu-Medrol 2 g) did not reverse the fulminant course of the respiratory failure and had no effect on pulmonary hypertension and the rightside heart failure. This patient is the last patient of our series, and now we routinely use prophylactic steroid (2 × 60 mg prednisone) in each patient previously exposed to BCNU, and we restrict the inspiratory oxygen concentration to less than 30% during and after surgery. In patients with diffusion capacity <60% in preoperative pulmonary function tests, we also give postoperative steroids for 14 days (20 mg every 12 hours). The high incidence of recurrence within the liver after curative resection (67%) indicates either that biologic factors, yet unknown, might predispose the liver as a favorable site of seeding in selected types of breast cancer or that micrometastases
Surgery April 2000
had remained undetected at the time of resection (67%). New imaging modalities such as the PET scan might be useful to enhance detection of small metastases throughout the liver.26 The use of selective adjuvant chemotherapy delivered to the liver with a subcutaneous pump device via a catheter inserted into the gastroduodenal artery may reduce the risk of early intrahepatic recurrence. Several prognostic factors were investigated in this series of 17 patients. The only significant factor identified was the time between primary breast surgery and the diagnosis of liver metastases. Patients in whom liver metastases developed within the first year after primary breast surgery had a significantly shorter survival (median survival, 9 months vs 27 months; P = .04 log rank test). Although aggressive preoperative treatment with ablative HDC was not associated with a statistically significant improvement in patient survival when compared with standard chemotherapy regimens, a benefit from HDC cannot be excluded. A bias in the selection of patients for preoperative therapy is unavoidable in this type of study. Similarly, although the lymph node status at the time of breast surgery had no significant influence on patient survival in this series, the role of initial staging of the breast cancer may be important. Both of the long-term survivors in our series had no lymph node involvement in the axilla. In summary, liver resection can be performed with minimal morbidity and low mortality in this patient group. Pulmonary toxicity from previous chemotherapy using BCNU should be carefully evaluated before surgery, and prophylactic steroids should be administered along with minimizing PO2. Liver resection in this selected group is associated with a favorable 22% 5-year survival rate. Patients with a long interval (more than 1 year) between breast cancer resection and liver metastases are most likely to benefit from surgery. Additional adjuvant therapies, such as selective adjuvant intra-arterial chemotherapy, should be investigated to decrease the high incidence of hepatic recurrence after resection. REFERENCES 1. Lorenz M, Wiesner J, Staib-Sebler E, Encke A. Regionale Therapie von Mammakarzinom-Lebermetastasen. Zentralbl Chir 1995;120:786-90. 2. Raab R, Nussbaum KT, Pichlmayr R. Lebermetastasen bei Mammacarcinom. Chirurg 1996;67:234-7. 3. Foster JH. Survival after liver resection for secondary tumors. Am J Surg 1978;135:389-93. 4. Elias D, Lasser P, Spielmann M, et al. Surgical and chemotherapeutic treatment of hepatic metastases from carcinoma of the breast. Surgery 1991;172:461-4.
Surgery Volume 127, Number 4 5. Pocard M, Salmon R-J. Resection hepatique pour metastase du cancer du sein: le concept de chirurgie adjuvante. Bull Cancer 1997;84:47-50. 6. Selzner M, Clavien P-A. Resection of liver tumors: special emphasis on neoadjuvant and adjuvant therapy. In Clavien P-A, ed. Malignant Liver Tumors. Malden, Mass: Blackwell Science; 1999:137-149. 7. Fong Y, Cohen A, Fortner J, et al. Liver resection for colorectal metastases. J Clin Oncol 1997;15:938-46. 8. Gayowski TJ, Iwatsuki S, Madariaga JR. Experience in hepatic resection for metastatic colorectal cancer: analysis of clinical and pathologic risk factors. Surgery 1994;116:703. 9. Selzner M, Clavien P-A. Liver tumors of rare and unknown origins. In Clavien P-A, ed. Malignant Liver Tumors. Malden, Mass: Blackwell Science; 1999:303-312. 10. Hoe A, Royle G, Taylor I. Breast cancer and liver metastases— incidence, diagnosis and outcome. J R Soc Med 1984;84:714-6. 11. Zinser J, Hortobagyi G, Buzdar A, et al. Clinical course of breast cancer patients with liver metastases. J Clin Oncol 1987;5:773-82. 12. Patanaphan V, Salazar M, Risco R. Breast cancer: metastatic patterns and their prognosis. South Med J 1988;81:1109-12. 13. O’Reilly S, Richards M, Rubens R. Liver metastases from breast cancer: the relationship between clinical, biochemical and pathological features and survival. Eur J Cancer 1990;26:574-7. 14. Zimmermann T, Padberg W, Kelm C, et al. Die lokoregionale Chemotherapie zur Behandlung von Lebermetastasen des Mammakarzinoms. Zentralbl Chir 1992;117:226-30. 15. Schneebaum S, Walker MJ, Young D, et al. The regional treatment of liver metastases from breast cancer. J Surg Oncol 1994;55:26-32. 16. Raab R, Nussbaum KT, Behrend M, Weimann A. Liver metastases of breast cancer—results of liver resection. Anticancer Res 1998;18(3C):2231-3.
Selzner et al 389
17. Stehlin JS, Ipolyi PD, Greeff PJ, et al. Treatment of cancer of the liver. Ann Surg 1988;208:23-35. 18. Szakacs JG, Szakacs JE, Karl RC. Surgical resection versus perfusion in the treatment of metastatic and primary liver tumors. Ann Clin Lab Sci 1990;20:245-56. 19. Wolf RF, Goodnight JE, Krag DE, Schneider PD. Results of resection and proposed guidelines for patient selection in instances of noncolorectal hepatic metastases. Surg Gynecol Obstet 1991;173:454-60. 20. Chap L, Shipner R, Levine M, et al. Pulmonary toxicity of high-dose chemotherapy for breast cancer: a non-invasive aproach to diagnosis and treatment. Bone Marrow Transplant 1997;20:1063-7. 21. Wilczynski S, Erasmus J, Petros W, et al. Delayed pulmonary toxicity syndrome following high-dose chemotherapy and bone marrow transplantation for breast cancer. Am J Resp Crit Care Med 1998;157:565-73. 22. Todd N, Peters W, Ost A, et al. Pulmonary drug toxicity in patients with primary breast cancer treated with high-dose combination chemotherapy and autologous bone marrow transplantation. Am Rev Respir Dis 1993;147:1264-70. 23. Cherniack R, Abrams J, Kalica A. Pulmonary disease associated with breast cancer therapy. Am J Respir Crit Care Med 1994;150:1169-73. 24. Demirer T, Weaver C, Buckner C, et al. High-dose cyclophosphamide, carmustine and etoposide followed by allogeneic bone marrow transplantation in ptients with lymphoid malignancies who received prior dose-limiting radiation therapy. J Clin Oncol 1995;13:596-602. 25. Smith A. The pulmonary toxicity of nitrosoureas. Pharmacol Ther 1989;41:443-60. 26. Bender H, Kirst J, Palmedo H, et al. Value of 18 fluorodeoxyglucose positron emission tomography in the staging of recurrent breast carcinoma. Anticancer Res 1997; 17(3B):1687-92.