Claudia Little

On the Use of Cause-Specific Failure and Conditional Failure Probabilities: Examples from Clinical Oncology Data

Abstract Nonparametric maximum likelihood estimation of the probability of failing from a particular cause by time t in the presence of other acting causes (i.e., the cause-specific failure probability) is discussed. A commonly used incorrect approach is to take 1 minus the Kaplan-Meier (KM) estimator (1 – KM), whereby patients who fail of extraneous causes are treated as censored observations. Examples showing the extent of bias in using the 1-KM approach are presented using clinical oncology data. This bias can be quite large if the data are uncensored or if a large percentage of patients fail from extraneous causes prior to the occurrence of failures from the cause of interest. Each cause-specific failure probability is mathematically defined as a function of all of the cause-specific hazards. Therefore, nonparametric estimates of the cause-specific failure probabilities may not be able to identify categorized covariate effects on the cause-specific hazards. These effects would be correctly identified ...

Prognostic significance of serum lactate dehydrogenase in malignant lymphoma

The pretreatment serum lactate dehydrogenase level (LDH) was the single most important prognostic variable in 30 patients with diffuse histiocytic lymphoma treated between January 1973 and January 1977 with a poly‐drug chemotherapy program called the cyclophosphamide L2 protocol at the Memorial Sloan‐Kettering Cancer Center. A highly significant difference was found between the survival patterns of patients with LDH levels of 500 U or less and those with LDH levels greater than 500 U. (Two‐year survival rates were 67% and 13%, respectively.) A similar trend was observed for 25 patients with diffuse, poorly differentiated lymphocytic lymphoma treated with the same protocol, although this difference was not statistically significant. (Corresponding two‐year survival rates were 74% and 33%, respectively.) The association of LDH level with survival was evident even after adjustment for other factors of potential prognostic significance. Pretreatment serum LDH determinations may provide a useful means of stratifying patient populations when comparing treatment programs for advanced stage non‐Hodgkins lymphoma. Cancer 46:139–143, 1980.

Treatment of advanced diffuse histiocytic lymphoma: an analysis of prognostic variables.

Sixty‐five patients with Stages III and IV diffuse histiocytic lymphoma (DHL) were treated with two different and successive combination chemotherapy protocols. Twenty‐seven patients were treated with the cyclophosphamide (CTX) L2 protocol, which included maintenance chemotherapy for three years. Thirty‐eight patients received the NHL‐3 program. Both protocols included radiotherapy (1350–4000 rad) to areas of initial bulky disease or persistent tumor, as well as central nervous system prophylaxis with intrathecal methotrexate or cytosine arabinoside in patients with bone marrow involvement. Two‐year survival rates were 44 and 56%, respectively, for the CTX‐L2 and NHL‐3 protocols. Of the 65 patients, 59 were evaluable for response to therapy. The CTX‐L2 produced a 58% total response (TR) rate, 39% complete (CR), and 19% partial (PR). The patients on NHL‐3 achieved a TR rate of 82%, 33% CR, and 48% PR. The difference in TR was significant (P = 0.05), but in CR was not. Prior chemotherapy (P = 0.077) and serum lactic dehydrogenase (LDH) level above 500 U/liter (P = 0.01) significantly lessened the chances for achievement of a CR. However, sex, age, the presence of systemic symptoms, stage (III vs. IV), and prior RT were not found to be significantly related to CR rate. This analysis suggests that a high level of serum LDH characterizes a subgroup of patients with particularly aggressive DHL that requires a more intensive modality of treatment.

Reasons that patients with acute myelogenous leukemia do not undergo allogeneic bone marrow transplantation.

BACKGROUND Numerous reports suggest that allogeneic bone marrow transplantation prolongs the survival of adult patients with acute myelogenous leukemia (AML) in first remission. However, it is unclear how many such patients actually undergo this procedure. METHODS We reviewed the case records of 350 consecutive adult patients with AML treated with chemotherapy at a single institution from 1979 (when the policy of offering allogeneic transplantation to all such patients in first remission was introduced) through 1990. The criteria for exclusion before transplantation included age greater than 40 and, beginning in 1984, a diagnosis of acute promyelocytic leukemia. RESULTS One hundred forty-two patients (41 percent of the study population) were 40 years of age or under. HLA testing was performed for 120 of these patients (85 percent). Sixty-seven patients (47 percent) had an HLA-identical sibling as a potential donor. One hundred three patients (73 percent) entered remission during treatment according to one of five chemotherapy protocols. Of the 52 patients who both entered remission and had an HLA match, 30 underwent transplantation while they were in first remission. These 30 patients constituted 21 percent of all study patients 40 years old or under, 29 percent of all patients 40 or under who entered remission, 45 percent of all patients with an HLA match, 58 percent of all patients who had both a remission and a match, and 9 percent of all patients treated according to a protocol. Among patients with a match who did not undergo transplantation, those with primary refractory disease were the largest subgroup. CONCLUSIONS These findings suggest that allogeneic bone marrow transplantation is performed in less than 60 percent of adult patients with AML who are potentially eligible for the procedure.

Importance of long-term follow-up in evaluating treatment regimens for adults with acute lymphoblastic leukemia.

During the past 20 years, we have treated 250 previously untreated adults (greater than age 15 years) with acute lymphoblastic leukemia (ALL) with five successive multidrug protocols: L2, L10, L10M, L17/17M, and L20. The L10 and L10M protocols had the highest percentage of long-term (greater than 5 years) remissions (52% and 40% respectively) compared with the L2 and more recent protocols (24%-32%); this is partly attributable to a greater prevalence of adverse risk factors among the latter protocols. The overall long-term survival of the first 199 patients with minimum 3 years follow-up is now 31%, with 35% of the 163 patients achieving complete remission (CR) remaining free of relapse for greater than 5 years. The disease-free survival of the 163 patients reaches a plateau of 33% after 6 years. The percentages of patients subsequently relapsing after remaining in continuous CR for 1.5, 3, and 5 years are 42%, 28%, and 6%, respectively; no relapses have yet occurred after 6 years in this series. Postrelapse survival improved progressively with longer duration of first remission. The results of treatment in second or later remission with either chemotherapy or bone marrow transplantation (BMT) were unsatisfactory and there were only a few long-term survivors. Recently we have attempted to select patients at highest risk of early relapse for BMT in first remission, but the number of eligible patients actually having BMTs has been low for a variety of reasons, including early death, failure to reach CR, early relapse, patient refusal, or medical contraindications.(ABSTRACT TRUNCATED AT 250 WORDS)

Current status of treatment of acute leukemia in adults: An overview of the memorial experience and review of literature

The results of treatment of 629 previously untreated adults with acute leukemia at Memorial Hospital are reviewed. During the past 14 years, 135 adults (greater than 15 years) with acute lymphoblastic leukemia (ALL) have been treated with one of three successive multidrug-intensive treatment protocols (L2, L10/10M, and L17/17M), each calling for 2.5 to 3 years of systemic chemotherapy and prophylactic intrathecal methotrexate without cranial irradiation. The complete remission (CR) rates were L2 (n = 22) = 77%; L10/10M (n = 69) = 86%; L17/17M (n = 44) = 77%. The median durations of survival and remission were, respectively, L2 = 33 and 30 months; L10/10M = 62 months and not reached; and L17/17M = not reached. Almost all relapses occurred within the first 3 years while still continuing treatment, and there were only rate late relapses after stopping treatment. It appears that approximately half of the patients may have been cured with the latest two protocols. During the last 17 years, 494 adults aged 15 to greater than 70 with acute nonlymphoblastic leukemia (ANLL) were treated with one of five successive multiple drug treatment protocols of varying intensity (arabinosylcytosine + 6-thioguanine [n = 36]; L6 [n = 101]; L12 [n = 104]; L14/14M [n = 121]; and L16/16M [n = 132]). Patients with myelodysplastic syndromes generally were not treated until they developed acute leukemia, but were then entered and included in the results. Secondary leukemias following treatment of other neoplastic diseases were not included. The complete remission rates were fairly constant between 47 and 64% and the median durations of remissions were between 9 and 21 months. The intensive treatment L14 and L16 protocols were associated with more early deaths and did not result in a significantly improved remission incidence or duration or survival. With all protocols, the majority of relapses occurred within the first 2 years, but relapses continued to occur at a decreasing rate for 4 years and occasionally even later. Whereas a small fraction (approximately 10 to 15%) of adults with ANLL are now apparently being cured with combination chemotherapy, despite intensive efforts there has been little improvement during the last decade and more selective and effective forms of treatment are urgently needed.

Treatment of Diffuse Poorly Differentiated Lymphocytic Lymphoma An Analysis of Prognostic Variables

Forty patients with diffuse poorly differentiated lymphocytic lymphoma (DPDL) Stages II‐IV were treated with three different and successive combination chemotherapy protocols. Seventeen patients were treated with the cyclophosphamide (CTX) L2 protocol which included maintenance chemotherapy for 3 years. Only three patients were treated with the NHL‐3 (non‐Hodgkins lymphoma) protocol, and 20 patients received the NHL‐5 program. All protocols included radiotherapy (1500–4800 rad) to areas of initial bulky disease or persistent tumor, as well as central nervous system (CNS) prophylaxis with intrathecal methotrexate in patients with bone marrow involvement. Seventy‐eight percent of local recurrences occurred in previously irradiated areas. Two‐year survival rates were 55% and 70% for the CTX‐L2 and NHL‐5 protocols, respectively. Median disease‐free survival for 24 complete response (CR) patients was 16.5 months. of the 40 patients, 37 were evaluable for response to therapy. The CTX‐L2 produced an 80% total response (TR) rate, a 60% CR, and a 20% partial response (PR). The patients on the NHL‐5 achieved a TR rate of 95%, 74% CR, and 21% PR. Differences in TR and CR between the two protocols were not significant. Only 1 of 3 patients on the NHL‐3 protocol achieved a CR. There was a trend for age >50 years to lessen the chances of CR (P = 0.091); however, sex, symptoms, stage of disease, and LDH level were not significantly related to CR rate. Response to treatment (CR versus PR versus failure) was the most important factor influencing survival (P <0.001); age (>50 years) was also significant (P = 0.008). Lactate dehydrogenase (LDH) was of borderline significance (P = 0.06). Cox regression model showed age (>50 versus <50 years, P = 0.001), LDH (>500 versus≦500 U/L, P = 0.019) and symptoms (A or B) to be the best predictors of survival.

NHL‐3 protocol six‐drug combination chemotherapy for non‐hodgkin's lymphoma

Combination chemotherapy and radiotherapy (RT) were administered to 73 adults with non‐Hodgkins lymphoma (NHL). Ten cycles of the following drugs were given: intravenous Adriamycin (doxorubicin) (25 mg/m2), cyclophosphamide (700 mg/m2) and vincristine (1.5 mg/m2) on day 1; arabinosylcytosine (100 mg/m2) and methotrexate (10 mg/m2) on days 3 to 5; and oral prednisone (60 mg/m2) on days 1 to 5. Radiotherapy was given to resistant or initially bulky disease (2000 rad). Patients were also randomized to receive pseudomonas vaccine or no immunotherapy. of 61 evaluable patients, 33 (54%) achieved a complete response (CR) and 18 (30%) a partial response (PR). Among 44 evaluable patients with diffuse histiocytic lymphoma (DHL), 22 (50%) had a CR, and 15 (34%) a PR. for 17 evaluable patients with nodular (4) and diffuse (11) mixed and poorly differentiated lymphocytic and diffuse “undifferentiated” (2) lymphomas, CR and PR rates were 65% and 18%, respectively. No statistically significant differences in response rate or duration and survival have been observed between the patients randomized to receive pseudomonas vaccine or no immunotherapy. Median follow‐up time from start of treatment was 47.5 months. Median survival for all 73 patients (including inevaluables) and for 52 DHL patients was 30.7 months. Poor prognostic features influencing survival included: female sex (P = 0.003), poor response to therapy (CR versus PR; P = 0.001), prior chemotherapy, (P = 0.01) and high levels of lactic dehydrogenase (P = 0.001). It can be concluded that this combination of cycle and phase‐active agents is of similar efficacy to other reported regimens in inducing major responses and that it has the potential to prolong disease‐free survival. The analysis of prognostic factors has been used to dissect poor prognostic categories that might require different modalities of treatment.

Financial analysis of patients with newly diagnosed acute myelogenous leukemia on protocol or standard therapy

Medicare and third‐party payers may be reluctant to pay for investigational (protocol) therapy for patients with cancer on the premise that such treatment is more expensive than standard therapy. However, prior studies that have attempted to compare protocol therapy with standard therapy have been difficult to interpret because of the assortment of malignancies studied and the lack of suitable control groups of patients who received standard therapy.