The natural history and patterns of metastases from mucosal melanoma: an analysis of 706 prospectively-followed patients.

Background: We examined whether mucosal melanomas are different in their clinical course and patterns of metastases when arising from different anatomic sites. Our hypothesis was that metastatic behavior would differ from primary mucosal melanomas at different anatomical sites. Patients and methods: Clinical and pathological data from 706 patients were compared for their stage distribution, patterns of metastases, CKIT/BRAF mutation status, and overall survival for different anatomical sites. Results: The anatomic sites of the primary mucosal melanomas were from the lower GI tract (26.5%), nasal cavity and paranasal sinuses (23%), gynecological sites (22.5%), oral cavity (15%), urological sites (5%), upper GI tract (5%), and other sites (3.0%). At initial diagnosis, 14.5% were stage I disease, 41% Stage II, 21.5% Stage III, and 23.0% stage IV. Predominant metastatic sites were regional lymph nodes (21.5%), lung (21%), liver (18.5%), and distant nodes (9%). Oral cavity mucosal melanoma had a higher incidence of regional nodal metastases (31.7% versus 19.8%, P = 0.009), and a higher incidence of lung metastases (32.5% versus 18.5%, P = 0.007) compared to other primary mucosal melanomas. There was a 10% incidence of CKIT mutation and 12% BRAF mutation. Mucosal melanomas from nasal pharyngeal and oral, gastrointestinal, gynecological, and urological had a similar survival with a 1-year survival rate (88%, 83%, 86%), 2-year survival rate (66%, 57%, 61%), 5-year survival rate (27%, 16%, 20%), respectively. Conclusions: The largest sample size allows, for the first time, a comparison of primary melanoma stage and patterns of metastases across anatomical sites. With few exceptions, the presenting stages, incidence of nodal and distant metastases, the site of predilection of distant metastases, or overall survival were similar despite different primary anatomic sites. These findings suggest that clinical trials involving mucosal melanomas and the administration of systemic therapy can be applied equally to mucosal melanomas regardless of their primary anatomic site.

Burnout and career satisfaction among surgical oncologists compared with other surgical specialties.

INTRODUCTION: Little is known regarding the rate of burnout, career satisfaction, and quality of life (QOL) among surgical oncologists compared with other surgical subspecialties. METHODS: The American College of Surgeons conducted a survey in 2008 involving 7,905 respondents, of whom 407 were surgical oncologists. Demographic variables, practice characteristics, career satisfaction, burnout, and quality of life (QOL) of surgical oncologists were compared with other surgical subspecialties using validated instruments. RESULTS: Surgical oncologists were younger (mean age 49.9 years), more likely to be female (26%), and had younger children than other surgical subspecialties. With respect to practice characteristics, surgical oncologists had been in practice fewer years and had fewer nights on call per week than other surgical disciplines but worked more hours (mean 62.6/week), were more likely to be in an academic practice (59.5%), were more likely to be paid on a salaried basis (68%), and had more time devoted to non-patient activities (e.g., research). Compared with surgeons from all other specialties, surgical oncologists had similar incidence of burnout (36%), suicide ideation (4.9%), and QOL, but lower incidence of depression (24%), and better indices of career satisfaction. CONCLUSIONS: These data provide a frame of reference for valid comparisons of burnout, QOL, and career satisfaction indices for the surgical oncology community relative to all other surgical specialties. Surgical oncologists have higher career satisfaction and lower risk of depression than surgeons in other surgical disciplines but still experience high rates of burnout.

Autologous tumor-specific cytotoxic T lymphocytes in the infiltrate of human metastatic melanomas. Activation by interleukin 2 and autologous tumor cells, and involvement of the T cell receptor.

TIL from metastatic melanoma proliferated by greater than 1,000-fold (840-3,675, mean 1,543) after 6 wk in culture of mixtures of TIL and tumor cells with rIL-2 alone. Cytolysis was restricted to autologous tumor cells. CD8+ T cells were the predominant population of TIL before and after expansion, and were primarily responsible for autologous tumor-specific CTL activity. No other rIL-2-activated lymphocytes from peripheral blood, lymph nodes with melanoma metastasis, or TIL from sarcoma or renal cell carcinoma had autologous tumor-specific CTL activity. There were few or no CD16+ NK cells in TIL from metastatic melanoma before or after incubation with rIL-2, respectively. However, TIL from sarcoma or renal cell carcinoma contained a substantial proportion of CD3-CD16+ NK cells, which increased in number in culture with rIL-2. Purified CD16+ NK cells as well as CD3+CD16- T cells from rIL-2-activated TIL of renal cell carcinoma displayed MHC-nonrestricted cytotoxicity. At the clonal level as determined by limiting dilution, 8 of 10 clones from melanoma TIL displayed cytotoxicity restricted to autologous tumor cells, while all 13 clones from renal cancer TIL equally lysed autologous and allogeneic tumor cells. Anti-T cell receptor (TCR)-alpha/beta(WT31) mAb as well as anti-CD3 mAb inhibited autologous melanoma cell-specific CTL activity mediated by rIL-2-activated TIL at the effector phase. These two mAbs also inhibited rIL-2-dependent proliferation of these TIL when added to the culture. Pretreatment of fresh melanoma cells with mAb to MHC antigens followed by washing inhibited specific CTL activity. These results suggest that both TCR-alpha/beta on effector TIL and MHC antigens on fresh tumor cells are involved in the specific immune-recognition. After reaching maximum propagation, TIL from metastatic melanoma responded poorly to rIL-2 alone. However, stimulation with fresh autologous melanoma cells restored both CTL activity and proliferation in response to rIL-2. The latter is associated with IL-2 receptor (Tac antigen) expression on the surface. These results indicate that TIL from metastatic melanomas may have unique characteristics different from lymphocytes obtained from the other sources, and may contain precursor CTL sensitized in vivo to autologous tumor cells, and thus can be propagated in larger numbers with rIL-2 alone while retaining autologous tumor-specific CTL activity.

Postnatal expansion of the natural killer and keller cell population in humans identified by the monoclonal HNK-1 antibody.

Human natural killer (NK) and killer (K) cells were directly enumerated using a monoclonal antibody (HNK-1) and an immunofluorescence assay. The frequency of cells bearing surface HNK-1 antigen was very low in the newborn (less than 1.0%) and increased progressively through childhood and into adult life. This was correlated with an age-related increase in functional NK and K cell activities. Males had a slightly higher proportion of HNK-1+ cells than females. In addition to HNK-1 expression on the surface membrane, a prominent cytoplasmic expression of HNK-1 antigen was found in some but not all surface HNK-1+ cells. The cytoplasmic accumulation of HNK-1 molecules appeared to occur in more mature cells of this lineage.

Differentiation stages of human natural killer cells in lymphoid tissues from fetal to adult life.

Virtually all human granular lymphocytes expressed the HNK-1 differentiation antigen when examined in lymphoid compartments from adults, neonates, and fetuses. The HNK-1+ cells were distinguishable into three subsets having distinct antigenic phenotypes: HNK+T3-M1-, HNK+T3+M1-, and HNK+T3-M1+. Thus, greater than 70% of the HNK-1+ cells from 13-17 wk fetuses (less than 0.2% of nucleated cells) lacked T cell antigens (e.g., T3, T8, T4, and T6) and the M1 myeloid antigen. Morphologically, the HNK+T3-M1- cells consisted of three different types: small granular lymphocytes (less than 10% of HNK-1+ cells), agranular small lymphocytes with a narrow rim of cytoplasm (70-80%), and agranular giant cells (greater than 15 micrometers) with considerable neutrophilic cytoplasm (15%). The purified fetal HNK-1+ cells exhibited a low level of cytotoxicity against K562 target cells. On the other hand, almost all of HNK-1+ cells in neonatal tissues as well as adult bone marrow, lymph node, and thymus, exhibited the HNK+T3+M1- phenotype, contained sparse cytoplasmic granules, and had an intermediate level of NK functional activity. Only adult blood and spleen contained a majority of mature HNK-1+ cells. These cells had an HNK+T3-M1- phenotype, abundant cytoplasmic granules, and maximum NK function. We propose that human NK cells may generate from a separate cell lineage and that they alter their phenotype, morphology, and functional capability during differentiation.

Thymus-dependent lymphocytes in tissue sections of rejecting rat renal allografts.

Lewis kidneys were grafted into BN recipients and examined at daily intervals up to 6 days after grafting with immunofluorescent reagents. A horse antiserum specific for T lymphocytes revealed an increasing number of T lymphocytes in the cellular infiltrates of rejecting allografts. These were detectable 1 day after grafting, reached a maximum 3 days later, and were relatively diminished at 6 days. In control isografts and nonimmunological inflammations of kidney, a small number of dispersed T lymphocytes was seen. A rabbit antirat thymocyte antiserum, given to allografted BN rats, prolonged survival of the grafts and decreased the cellular infiltrate and the number of T lymphocytes in the infiltrates. We conclude that in graft rejection there is a flow of T lymphocytes into areas of tissue damage and these T lymphocytes are immunologically reactive to graft antigens.

Isolation and partial characterization of the human homologue of Thy-1.

A glycoprotein of 25,000 daltons was isolated from a human T lymphoblastoid cell line (MOLT-3). A monovalent antiserum raised against this antigen precipitated mouse Thy-1.1 antigen as well as a 25,000-dalton antigen from human blood T cells. The unabsorbed serum was also cytotoxic for both mouse and human T cells. By multiple criteria, this human T antigen appears to be homologous with mouse Thy-1.

Ethnic differences in the lymphocyte proliferative response induced by a murine IgG1 antibody, Leu-4, to the T3 molecule.

The mitogenic effects of isotypically diverse antibodies to the T3 molecule were examined in genetically diverse population groups. Whereas the OKT3 antibody (IgG2a) was mitogenic for blood mononuclear cells from all individuals tested, the 38.1 antibody (IgM) was consistently nonmitogenic. In contrast, studies of the mitogenic effects of the Leu-4 antibody (IgG1) revealed striking ethnic differences. More than 80% of Caucasians and Negroes were good Leu-4 responders, whereas most individuals of Asian origin, including Indian, Japanese, and Chinese, were either Leu-4 nonresponders or Leu-4 low responders. However, the majority of American Indians, as well as a significant minority of Chinese, were good responders. Cell separation studies confirmed that monocytes govern the different mitogenic effects of the anti-T3 antibodies. The results reveal interesting ethnic differences in monocyte accessory function probably mediated via the Fc-gamma receptor, in the stimulation of T lymphocytes by an IgG1 antibody against the T3 molecule.

How pathologic staging informs the surgical management of primary cutaneous melanoma.

In this brief review, the authors will summarize recent changes in the American Joint Committee on Cancer (AJCC) staging system, with a particular emphasis on the prognostic importance of mitotic rate and microsatellitosis, and address continuing controversial aspects of the surgical management of cutaneous melanoma.

Natural killer (HNK-1+) cells in Chediak-Higashi patients are present in normal numbers but are abnormal in function and morphology.

Children with the Chediak-Higashi (CH) syndrome are known to have abnormalities of natural killer (NK) cell function. We used the HNK-1 monoclonal antibody that reacts specifically with human NK and K cells to distinguish whether this abnormality was due either to a numerical deficiency of NK cells or a defect in their ability to function. In eight CH patients, a significant proportion of their blood mononuclear cells (10--19%) expressed the HNK-1 differentiation antigen. The level of NK cells in the five children with CH syndrome was higher than for age-matched normal controls (15.8% vs. 5.8%, P less than 0.001). When HNK-1+ cells were isolated with a fluorescence-activated cell sorter, the NK cells from CH patients were a homogeneous population of lymphocytes with a single large granule rather than the multiple small granules seen in Nk cells from normal individuals. The purified HNK-1+ cells from the CH patients had minimal NK or K cell function. The CH syndrome thus includes a functionally defective population of NK cells that retain the capability of expressing the HNK-1 differentiation antigen.

Lack of comparability between binding of monoclonal antibodies to melanoma cells in vitro and localization in vivo.

The affinity and specificity of monoclonal antibodies (MAbs) to tumor-associated antigens are often determined by in vitro assays. The specific binding of two anti-human melanoma antibodies (96.5 and ZME-018) and a control antibody (ZCE-025) to three human melanoma cell lines (DX3, A375-M, and Hs294t) was examined under in vitro conditions and compared to in vivo localization of 111In-labeled antibodies to the same cells growing as solid tumors in the subcutis of nude mice. The in vitro binding of the specific MAbs to the tumor cells did not predict in vivo localization. Under in vitro conditions, MAb ZME-018 bound to all three cell lines at levels exceeding that of 96.5, yet ZME-018 did not show superior localization to subcutaneous tumors. MAb 96.5 bound to cultured DX3 cells at levels exceeding those observed with A375-M cells. Yet, 96.5 localized better to A375-M xenografts in nude mice than to DX3 or Hs294t xenografts. Antigen expression differed between in vitro and in vivo growing cells, as evidenced by alteration in binding of 96.5 to tumor cells dissociated from solid subcutaneous tumors. Collectively, the data suggest that in vitro parameters do not predict the clinically relevant localization of MAbs to tumors.

Interleukin 2 activation of cytotoxic T-lymphocytes infiltrating into human metastatic melanomas.

Tumor infiltrating lymphocytes (TIL) were isolated from 22 tumors obtained from 15 patients with metastatic melanoma. In 18 of the 22 tumors, a substantial number of lymphocytes was isolated with an average lymphoid cell:tumor cell ratio of 1.26. The TIL were predominantly cytotoxic/suppressor T-lymphocytes with an average of 87% Leu4+, 61% Leu2a+, and 18% Leu3a+ cells. There were less than 2% natural killer cells, B-cells, or macrophages. An average of 3.8% (range, less than 0.1 to 8.6%) of freshly isolated TIL bound to autologous tumor cells. Prior to culture, none of the tumor-binding cells (TBC) was cytotoxic as judged by trypan blue exclusion. The frequency of TBC increased to 11.6% after 2 days of culture, and 10% of these TBC developed cytotoxic activity. When interleukin 2 was added to cultures, the frequency of TBC increased, and the frequency of cytotoxic TBC was 2-fold higher compared to control cultures. After 10 days of culture with interleukin 2, TIL increased in number with a concomitant disappearance of tumor cells, whereas there were severe decreases of lymphocytes and no decrease of tumor cells in control cultures. TIL were cultured for 8 to 10 days with recombinant interleukin 2 and tested for cytotoxicity against autologous and allogenic tumor cells and K562 targets in a 4-h 51Cr release assay. rIL2-cultured TIL from all nine patients tested exhibited the highest levels of lysis against autologous tumor cells. Of the nine TIL samples, five exhibited an apparent specificity for autologous melanoma, while four specimens killed both allogenic and autologous melanoma. The ability of TIL to kill K562 targets appeared to parallel the ability to kill allogenic targets. For comparison, recombinant interleukin 2-cultured peripheral blood mononuclear cells from the same patients were assayed for cytotoxic activity against autologous and allogenic melanomas. Unlike some TIL, none of the peripheral blood mononuclear cells exhibited specificity for autologous tumor cells. In summary, TIL isolated from metastatic melanoma patients were predominantly cytotoxic T-lymphocytes with the ability to recognize and kill autologous tumor cells after in vitro culture; interleukin 2 induced proliferation of TIL and augmented their cytotoxic activity such that they eliminated autologous tumor cells.

In vitro migration of lymphocytes through collagen matrix: arrested locomotion in tumor-infiltrating lymphocytes.

Antitumor immunity requires (a) extravasation of lymphocytes from the blood stream to interstitium, (b) locomotion through extracellular matrix to the site of the tumor, (c) effector cell recognition of the tumor target with cell/cell contact and binding of adhesion receptors, (d) T-cell receptor binding to histocompatibility and tumor antigens, and (e) tumor cell lysis. We hypothesize that the tumor microenvironment inhibits lymphocyte locomotion through extracellular matrix as one mechanism by which tumors may avert host defense. Lymphocyte locomotion was investigated in vitro using a three-dimensional collagen gel model. Fresh tumor-infiltrating lymphocytes (TIL) were obtained by enzymatic digestion of melanomas and renal cell carcinoma, and mononuclear cells were isolated by discontinuous Ficoll-Hypaque gradient. The lymphocytes were analyzed for motility from a point of origin between basal and overlay layers of collagen gel. Results showed that TIL migration was almost completely inhibited, compared with migration of normal and cancer patient peripheral blood leukocytes and lymphocytes from lymph nodes. Short-term (24-h) exposure of lymphocytes to cytokines during the assay in the collagen gel matrix had no effect on locomotor ability. Long-term (19, 30, or 35 days) culture of TIL in 200 units/ml of interleukin 2 reinstated locomotor ability. Short-term exposure of any of the lymphocyte populations to interleukin 1-alpha, interleukin 1-beta, interleukin 2, interleukin 3, interleukin 4, alpha-interferon, or gamma-interferon had no effect on migration. Thus, TIL display a uniquely arrested ability to locomote through collagen gel. Inhibition of the locomotion of infiltrating effector cells is possibly a mechanism by which the tumor evades the host immune system.

Role of interleukin 2 and a serum suppressive factor on the induction of activated killer cells cytotoxic for autologous human melanoma cells.

Activated killer (AK) cells cytotoxic for freshly prepared autologous melanoma cells were generated in vitro when recombinant interleukin 2 (rIL2) was incubated with peripheral blood mononuclear cells from 11 of 12 patients with metastatic melanoma. The cytotoxic response first appeared significant after 3 days of culture with rIL2; it peaked after 6 days and then declined after 9 days of incubation. Peripheral blood mononuclear cells cultured with medium alone (control) or with recombinant gamma-interferon alone (200 units/ml) failed to acquire AK cytotoxicity at any time in any of the 12 patients. Autologous tumor cells also could not induce AK activity in mixed-lymphocyte tumor cell cultures. The level of AK activity generated was significantly reduced (P less than 0.001) when 10% autologous human serum was used instead of 10% fetal calf serum in the rIL2 cultures. This suppression was specific for the inductive phase of AK activity, because autologous human serum could not suppress the cytotoxic capability of AK cells once they were activated by rIL2. The serum suppressive effect on the induction of AK cytotoxicity could be overcome by increasing the dose of rIL2 or by adding recombinant gamma-interferon to the cultures. Lymphocytes from melanoma patients thus have the latent ability to kill autologous melanoma tumor cells. However, this cytotoxic capability requires an interleukin 2-induced differentiation step that could be amplified by gamma-interferon and inhibited by a serum suppressive factor.

A methodological approach to the prediction of anticancer drug effect in humans.

Tumor cells from animals and humans were treated with drugs under tissue culture conditions. Tumor cells from the sensitive L1210 model were studied first. A dose-response curve was derived between drug exposure and subsequent cytotoxicity in L1210. The concentration of drug and duration of exposure were factors critical to the subsequent development of in vitro cytotoxicity. The in vitro dosage which effected 50% leukemic cell death in L1210 cells correlated with reported in vivo drug levels. Other tumor models and human neoplastic cells were studied at this dosage level. A good correlation was noted in these studies between the in vivo responsiveness and the in vitro chemotherapy results in both animals and humans. It was suggested by these results that it may be possible to predict cancericidal drug activity for individual neoplasms by assaying the tumor cells in vitro for drug sensitivity.

Clonal analysis of lymphocytes from tumor, peripheral blood, and nontumorous kidney in primary renal cell carcinoma.

The immunological properties of lymphocytes from tumor, peripheral blood (PBL), and nontumorous kidney from 16 patients with renal cell carcinoma were characterized at the clonal level with respect to their clonogenic efficiency, phenotypic expression, and cytotoxicity against autologous and allogenic tumor cells. The objectives were to delineate: (a) the quantitative differences in the immunological properties of tumor-infiltrating lymphocytes (TIL) from patient to patient; and (b) the qualitative differences in immunological properties between TIL and lymphocytes from peripheral blood or nontumorous kidney from a single patient. A total of 926 clones were characterized for phenotype expression, and 465 clones were characterized for cytotoxicity. The clonogenic efficiency of TIL varied with individuals: high in one patient; relatively high to moderate in seven patients; low in seven patients, and extremely low in the remaining one patient. The levels of autologous tumor cell lysis by TIL clones also varied with individuals. More than one-third of the TIL clones established in 4 of 13 patients displayed significant (greater than or equal to 10%) lysis against autologous tumor cells, and in each of the four patients the average percentage of lysis in the total TIL clones was higher than 10%. In two patients, 5 of 26 or 3 of 13 TIL clones were cytotoxic, but averages of percentage of lysis in the total clones were less than 10%. One 1 or 2 TIL clones of 10-27 total clones were cytotoxic in each of 4 patients, while no cytotoxic TIL clones were found in the remaining 3 patients. Clonogenic efficiency did not correlate with the level of cytotoxicity, and TIL from no tumors displayed both high proliferation and high cytotoxicity at the clonal level. In a majority of patients (12 of 13), most cytotoxic TIL clones against autologous tumor cells also lysed allogenic tumor cells. In contrast, TIL clones lysed only autologous tumor cells in the remaining one patient (patient 2). The clonogenic efficiency of TIL was lower than that of PBL in 6 of 12 patients, while the opposite was true in the remaining 6 patients. The level of cytotoxicity in the PBL clones of these 12 patients primarily correlated with that of the TIL clones. With one exception (patient 2), most cytotoxic PBL clones against autologous tumor cells also lysed allogenic targets in a majority of patients. CD4+CD8-T-cell clones (70-85%) predominated in all patients regardless of the different lymphocyte sources.(ABSTRACT TRUNCATED AT 400 WORDS)

Involvement of both Tac and non-Tac interleukin 2-binding peptides in the interleukin 2-dependent proliferation of human tumor-infiltrating lymphocytes.

Interleukin 2 (IL-2) receptor expression was examined on recombinant IL-2 (rIL-2)-propagated tumor-infiltrating lymphocytes (TIL) from eight metastatic melanoma and three sarcoma samples. All 11 TIL expanded with similar growth rates. rIL-2 propagated TIL from five of eight metastatic melanoma specimens contained no Tac antigen-positive lymphocytes as determined by immunofluorescence and flow cytometry performed multiple times during the 4 to 8 week culture period. However, "Tac-negative" TIL did express the non-Tac IL-2-binding peptide, p70-75 as determined by [125I]IL-2 cross-linking and sodium dodecyl sulfate-polyacrylamide gel electrophoresis. IL-2-binding assays revealed that these "Tac-negative" TIL expressed only an intermediate affinity IL-2 receptor. In contrast, TIL from the other three of eight melanoma and all three sarcoma contained one-third Tac-positive cells as assessed by flow cytometry analysis, and expressed surface non-Tac (p70-75) and Tac (p55) peptides by [125I]IL-2 cross-linking. These "Tac-positive" TIL displayed both the high and intermediate affinity IL-2 receptors. However, rIL-2-dependent growth of both "Tac-negative" and "Tac-positive" TIL was significantly inhibited by anti-Tac mAb, suggesting a transient Tac expression on the "Tac-negative" TIL. Additionally, due to the limits of our methodology, we cannot rule out the possibility of a constitutive expression of a low level of Tac, with an indicible expression of higher levels. Addition of culture supernatants from phytohemagglutinin- and phorbol myristate acetate-stimulated peripheral blood mononuclear cells to the "Tac-negative" TIL-induced detectable Tac expression within 48 h. These results indicate that both non-Tac and Tac IL-2 receptors play important roles during IL-2-dependent proliferation of TIL.

The role of elective lymph node dissection in melanoma: rationale, results, and controversies.

Elective lymph node dissection (ELND) for patients with clinically occult metastatic melanoma in regional lymph nodes has the goal of curing metastases with a surgical treatment. This is in contrast to the low probability for surgical cure in patients with clinically detectable lymph node metastases. The rationale for elective node dissection is based on a hypothesis that melanoma metastasizes sequentially via lymph nodes and then to distant sites. A subgroup of melanoma patients with high risk for regional node micrometastases but low risk for distant micrometastases has been identified from prognostic factors analysis of large patient series, as well as surgical results of nonrandomized clinical trials. However, two nonrandomized surgical trials have failed to show a survival benefit for ELND. These studies were largely performed in female patients with extremity melanomas and there were limitations that preclude a definitive conclusion. No randomized trials have been conducted involving melanomas of the trunk or head and neck. Two prospective randomized surgical trials are now being conducted in North America and in Europe. Until the results of these trials are available, physicians are encouraged to enter patients into these ongoing clinical trials or consider ELND in selected patients where the benefit-risk ratio justifies it. Factors to be considered in this decision include intermediate tumor thickness (ie, 1 to 4 mm thickness), anatomic site, histology (ulceration and growth pattern), and the risk of the operation in individual patient settings.

Corynebacterium parvum versus bacille Calmette-Guérin adjuvant immunotherapy of stage II malignant melanoma.

Two separate studies have been reported comparing Corynebacterium parvum and bacille Calmette-Guérin (BCG) as adjuvant immunotherapy for stage II melanoma patients (The Milton S. Hershey Medical Center, 48 patients; Southeastern Cancer Study Group [SECSG], 162 patients). As the criteria for patient selection and drugs used were similar, we have pooled the data to analyze the effects of these two treatments. Both studies used BCG (Tice, Chicago, IL) 3 x 10(8) live organisms per treatment by Tine technique and C parvum (Burroughs-Wellcome, Triangle Park, NC) subcutaneous at a dose of 4 mg/m2 (SECSG) or 5 micrograms/m2 (Hershey) per treatment. The only difference in these studies was the frequency of immunization, with patients in Hershey receiving 22 doses and the SECSG patients receiving 55 doses during the 2-year period of treatment. Kaplan-Meier life-table analysis for the 210 patients shows a prolonged disease-free interval for patients treated with C parvum (P = .02, two-sided Mantel procedure). In similar fashion, patients treated with C parvum had an improved survival rate (from all causes) when compared with BCG-treated patients (P = .012). An analysis of the results for the 170 patients for which the number of positive nodes was available was performed using Cox's model, with nodes as a stratification variable and with covariates of place, treatment, age, and sex. In this analysis, an observed benefit for C parvum on the disease-free interval had a P value of .37 while the benefit of C parvum on the survival times (from all causes) had a P value of .04. When the same analysis was performed using only patients aged younger than 60 years, the observed benefit of C parvum on disease-free interval had a P value of .08 and the benefit of C parvum on survival times (from all causes) had a P value of .008.

A multifactorial analysis of prognostic factors in patients with liver metastases from colorectal carcinoma.

A multifactorial analysis was used to identify the dominant prognostic variables predicting survival rates of 175 patients with hepatic metastases from colorectal carcinoma. Seven of 22 parameters examined simultaneously were found to independently influence the median survival rate in these patients: (1) elevated alkaline phosphatase (p = 0.0004), (2) elevated serum bilirubin level (p = 0.0005), (3) location of hepatic metastases (unilateral or bilateral, p = 0.0022), (4) number of metastatic nodes involved (0, 1-5, greater than 5; p = 0.0148), (5) depressed serum albumin (p = 0.0217), (6) whether or not the primary colorectal tumor was resected (p = 0.0013), and (7) chemotherapy (given or withheld, p = 0.0439). The prothrombin time, serum lactic dehydrogenase, and the number of hepatic metastases also correlated with survival, but they did not independently predict survival rates after other more dominant factors were accounted for. A mathematical equation for predicting an individual patient's clinical course once they developed hepatic metastases was derived from this statistical analysis. In addition, a simple and clinically useful guide for predicting outcome was developed that integrated the two most important risk factors, alkaline phosphatase and bilirubin.