The authors conducted a retrospective review of the charts of 56 patients who underwent resection for multiple brain metastases. Of these, 30 had one or more lesions left unresected (Group A) and 26 underwent resection of all lesions (Group B). Twenty-six other patients with a single metastasis who underwent resection (Group C) were selected to match Group B by type of primary tumor, time from first diagnosis of cancer to diagnosis of brain metastases, and presence or absence of systemic cancer at the time of surgery. Statistical analysis indicated that Groups A and B were also homogeneous for these prognostic indicators. Median survival duration was 6 months for Group A, 14 months for Group B, and 14 months for Group C. There was a statistically significant difference in survival time between Groups A and B (p = 0.003) and Groups A and C (p = 0.012) but not between Groups B and C (p > 0.5). Brain metastasis recurred in 31% of patients in Group B and in 35% of those in Group C; this difference was not significant (p > 0.5). Symptoms improved after surgery in 65% of patients in Group A, 83% in Group B, and 84% in Group C. Symptoms worsened in 13% of patients in Group A, 6% in Group B, and 0% in Group C. Groups A, B, and C had complication rates per craniotomy of 8%, 9%, and 8%, and 30-day mortality rates of 3%, 4%, and 0%, respectively. Guidelines for management of patients with multiple brain metastases are discussed. The authors conclude that surgical removal of all lesions in selected patients with multiple brain metastases results in significantly increased survival time and gives a prognosis similar to that of patients undergoing surgery for a single metastasis.

Results of reoperation in 48 patients who developed recurrent brain metastases between January 1984 and April 1993 are presented. Median time from first craniotomy to diagnosis of recurrence (time to recurrence) was 6.7 months. Median Karnofsky performance scale (KPS) score prior to reoperation was 80. Recurrence was local in 30 patients, distant in 16 patients, and both local and distant in two patients. Median survival time after reoperation was 11.5 months. There were no operative mortalities. Multivariate analysis revealed that presence of systemic disease (p = 0.008), KPS scores less than or equal to 70 (p = 0.008), time to recurrence of less than 4 months (p = 0.008), age greater than or equal to 40 years (p = 0.51), and primary tumor type of breast or melanoma (p = 0.028) negatively affected patient survival time. These five factors were used to develop a grading system (Grades I-IV). Patients categorized in Grade I had a 5-year survival rate of 57%, whereas the median survival time of patients in Grades II, III, and IV was 13.4, 6.8, and 3.4 months, respectively (p < 0.0001). Overall, 26 patients developed a second recurrence after reoperation. Seventeen patients underwent a second reoperation, whereas nine did not. Patients undergoing a second reoperation survived a median of 8.6 additional months versus 2.8 months for those who did not (p < 0.0001). This study concludes that reoperation for recurrent brain metastasis can prolong survival and improve quality of life. A second reoperation can also increase survival. Five factors influence survival: status of systemic disease, KPS score, time to recurrence, age, and type of primary tumor. The grading system using these five factors correlates with survival time. Reoperation should be approached with caution in Grade IV patients because of their poor prognosis.

Although radical resection is the best treatment for malignant sacral tumors, total sacrectomy for such tumors has been performed in only a few instances. Total sacral resection requires reconstruction of the pelvic ring plus establishment of a bilateral union between the lumbar spine and iliac bone. This technique is illustrated in two patients harboring large, painful, sacral giant-cell tumors that were unresponsive to prior treatment. These patients were treated with complete en bloc resection of the sacrum and complex iliolumbar reconstruction/stabilization and fusion. Surgery was performed in two stages, the first consisting of a midline celiotomy, dissection of visceral/neural structures, and ligation of internal iliac vessels, followed by an anterior L5-S1 discectomy. The second stage consisted of mobilization of an inferiorly based myocutaneous rectus abdominis pedicle flap for wound closure, followed by an L-5 laminectomy, bilateral L-5 foraminotomy, ligation of the thecal sac, division of sacral nerve roots, and transection of the ilia lateral to the tumor and sacroiliac joints. Placement of the instrumentation required segmental fixation of the lumbar spine from L-3 down by means of pedicle screws and the establishment of a bilateral liaison between the lumbar spine and the ilia by using the Galveston L-rod technique. The pelvic ring was then reestablished by means of a threaded rod connecting left and right ilia. Both autologous (posterior iliac crest) and allograft bone were used for fusion, and a tibial allograft strut was placed between the remaining ilia. The patients were immobilized for 8 weeks postoperatively and underwent progressive rehabilitation. At the 1-year follow-up review, one patient could walk unassisted, and the other ambulated independently using a cane. Both patients controlled bowel function satisfactorily with laxatives and diet and could maintain continence but required self-catheterization for bladder emptying. The authors conclude that in selected patients, total sacrectomy represents an acceptable surgical procedure that can offer not only effective local pain control, but also a potential cure, while preserving satisfactory ambulatory capacity and neurological function.

One‐hundred‐two patients with malignant melanoma who had distant metastases surgically resected and were judged to be clinically free of disease (M. D. Anderson Stage IVA melanoma) were studied. The median survival for all the patients from time of diagnosis of stage IVA disease was 18 months. The site of the resected metastases did not appear to influence survival, being approximately the same for the brain (15 months), lung (16 months), intraabdominal (18 months), and skin and/or lymph nodes (23 months). The site of the resected metastases also did not influence the median disease‐free interval. Patients who had metastases resected from several organs at the same time had a median survival of 15 months, which was similar to that of patients with one resected site. Patients who were rendered Stage IVA on several occasions by surgical excisions had a median survival of 36 months. Thirty‐five patients received surgery only and 67 patients received adjuvant chemotherapy, immunotherapy, or combined chemoimmunotherapy after surgery. For the group treated with surgery only, the median disease‐free interval and survival from diagnosis of stage IVA disease were 6 months and 16 months, respectively, and for the adjuvant group 6 months and 21 months, respectively. Specifically, by the type of adjuvant therapy, the median disease‐free interval and survival from stage IVA for 23 patients receiving Corynebacterium parvum were 6.9 and 19 months; for 39 patients receiving BCG, eight months and 26 months; for 24 patients receiving BCG + DTIC, eight and 17.4 months; and for all 51 DTIC treated patients 6.3 and 17.8 months, respectively. Patients receiving BCG had a median survival superior to the surgery only group (P = 0.02). An increase in survival was seen predominantly in patients who achieved IVA status more than once and received BCG. Patients with recurrent soft‐tissue metastases appeared to benefit most from BCG in prolonging the disease‐free interval. Only 1/10 treated by surgery alone had a disease‐free interval longer than 1 year, compared with 9/16 who received BCG (P = 0.01). Stage IVA melanoma appears to be distinctly different in prognosis from Stage IVB melanoma and should be classified separately. Patients with recurrent soft‐tissue disease may benefit significantly from treatment with BCG.