Jeffrey J. Gaynor

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 ...

Patterns of failure following surgical resection of renal cell carcinoma: implications for adjuvant local and systemic therapy.

PURPOSE This report is a patterns-of-failure analysis of resected renal cell carcinoma (RCC) performed to determine the relative incidences of local failure (LF) and distant failure, to identify the pathologic features predicting for each using a multivariate analysis, and to assess the relative impact of each form of failure on overall survival (OS). In this way, the potential value of and selection of patients for adjuvant local and/or systemic therapy can be better evaluated. MATERIALS AND METHODS The records of 172 patients with unilateral, nonmetastatic RCC who were treated with definitive surgery between 1978 and 1988, and who had a minimum follow-up duration of 1 year, were identified through the Memorial Sloan-Kettering tumor registry. Distribution by stage included T1, 10 patients; T2, 102; T3a, 32; T3b, 27; and T4, one. The incidences of positive lymph nodes (LNs) and positive margins were 5.8% and 6.4%, respectively. RESULTS LF developed in only six patients, yielding a 7-year actuarial incidence of 5%. In this subset, four patients developed distant metastases (DM), three occurring concurrently with or before LF. DM developed in 30 patients, yielding a 7-year actuarial incidence of 26%. Among the variables that had an impact on the development of DM according to univariate log-rank tests, only positive LNs (P = .026) and renal vein extension (P = .001) remained as significant independent prognosticators. The overall 7-year actuarial survival rate was 80%. Eleven patients died of RCC during follow-up, nine of whom (82%) died of metastatic disease. CONCLUSION LF is rare following surgical management of RCC, and shows no clear causal relationship with the development of DM. Patients die of DM, and not LF. These data do not support the role of adjuvant radiation therapy in this disease. Patients with LN involvement or renal vein extension have a significantly increased risk for developing DM, and are therefore appropriate candidates for trials investigating systemic therapy.

Long-term results of a prospective randomized trial of adjuvant brachytherapy in the management of completely resected soft tissue sarcomas of the extremity and superficial trunk☆

PURPOSE A prospective randomized trial evaluating adjuvant brachytherapy for soft tissue sarcomas was conducted between July 1982 and July 1987. METHODS AND MATERIALS One hundred twenty-six patients with soft tissue sarcoma of the extremity or superficial trunk (STS) underwent grossly complete resection with limb-sparing surgery. Intra-operatively, patients were randomized to receive either adjuvant brachytherapy (BRT) or no further therapy (no BRT). BRT consisted of an Iridium-192 implant which delivered 4200-4500 cGy over 4-6 days. Total hospital stay for combined surgery and BRT was 10-14 days. Patients in each group were well matched with respect to age, sex, site, tumor size, depth, histologic type, and grade. Median follow-up is 66.5 months for all living patients. RESULTS At 5 years, local control was 82% in the BRT group vs. 67% in the no BRT group (p = .049). When analyzed by histologic grade, high grade tumors had local control of 90% with BRT vs. 65% with no BRT (p = .013). There was no difference in local control in the low grade patients in either arm. At 5 years, the proportion free of distant metastases was approximately 76% in both arms, with no difference between BRT and no BRT. When analyzed by grade, high grade patients had a similar proportion free of distant metastasis in the BRT vs. no BRT arms despite improved local control in those receiving BRT. Similarly, the disease-specific survival for all patients at 5 years was 81% for BRT vs. 80% for no BRT (p = NS). When analyzed by grade, and treatment (BRT vs. no BRT), 5-year disease-specific survival for high grade was the same in both groups. CONCLUSIONS Adjuvant BRT significantly improves local control in completely resected STS. This improvement is confined to patients with high grade histology. BRT provides excellent local control in this group, and significantly shortens the treatment time when compared to adjuvant external beam radiation. However, the improved local control does not translate into either decreased distant metastasis or increased disease-specific survival.

Prognostic factors among 185 adults with newly diagnosed advanced Hodgkin's disease treated with alternating potentially noncross-resistant chemotherapy and intermediate-dose radiation therapy.

The initial promising results with alternating chemotherapy regimens (mechlorethamine, vincristine, procarbazine, and prednisone/doxorubicin, bleomycin, vinblastine, and dacarbazine [MOPP/ABVD]; lomustine, melphalan, and vindesine [CAD] plus MOPP plus ABV) combined with intermediate-dose radiation therapy (RT) have been sustained with further follow-up; 82.2% of patients (152 of 185) achieved a complete remission (CR), and overall survival is 71.7% +/- 4.4% at 8 years (median follow-up is 55 months among the survivors). No statistically significant differences were found in CR percentage, CR duration, or survival between stages IIB, IIIB, and IV patients. For that reason, stepwise Cox regression analyses to identify the important prognostic factors were performed on overall survival, tumor mortality, freedom from disease progression, and survival following disease progression. Pretreatment characteristics were also tested for association with the probability of achieving CR, CR duration, and death due to other causes. Characteristics that were consistently associated with an independently unfavorable prognosis were low hematocrit, high serum lactic acid dehydrogenase (LDH), age more than 45 years, inguinal node involvement, mediastinal mass greater than .45 of the thoracic diameter, and bone marrow involvement. Patients with two or more unfavorable characteristics were much more likely to fail treatment (median survival, 62.4 months) than those with none or only one unfavorable factor (greater than 95% survival). This striking difference between the low- and high-risk groups remained even if the comparison was restricted to patients less than or equal to 45 years of age. These results provide a basis for selecting the young patients at high risk of failure for more intensive initial treatment with either autologous bone marrow rescue or hematopoietic growth factors.

Performance status after treatment for squamous cell cancer of the base of tongue--a comparison of primary radiation therapy versus primary surgery.

PURPOSE To compare the quality of life and functional outcome in patients with squamous cell cancer of the base of tongue treated with primary radiation vs. primary surgery. METHODS AND MATERIALS At our institution, patients with base of tongue cancer are primarily treated either by radiation or surgery depending upon the philosophy of their primary physician. Primary radiation consists of 45-54 Gy external beam radiation followed by an 192Ir implant delivering an additional 20-30 Gy over 2-3 days. A neck dissection is done at the same time as the implant for those with involved nodes. Primary surgery consists of resection of the base of tongue lesion, neck dissection and postoperative radiation therapy. Because both groups have similar local control in our experience (80-90%), we used a subjective performance status scale for head and neck cancer patients to assess the quality of life in these patients (0-100, 0 = worst function, 100 = normal function). This scale measures ability to eat in public, understandability of speech, and normalcy of diet. There were 30 radiation patients (21: T1-T2; nine: T3-T4) and ten surgery patients (five: T1-T2; five: T3-T4) available for long-term quality of life assessment. RESULTS Patients treated with radiation had consistently better performance status scores and quality of life according to our study. This was true for those with early (T1-2) as well as more advanced (T3-4) disease. For eating in public, T1-2 patients had scores of 85 vs. 75 (p = .31) and T3-4 patients had scores of 82 vs. 35 (p < .0001) for radiation vs. surgery, respectively. For understandability of speech, T1-2 patients had scores of 92 vs. 65 (p = .0021), and T3-4 patients had scores of 95 vs. 35 (p < .0001) for radiation vs. surgery, respectively. For normalcy of diet, T1-2 patients had scores of 74 vs. 50 (p = .047), and T3-4 patients had scores of 78 vs. 32 (p = .0012) for radiation vs. surgery, respectively. In addition, we compared scores for early vs. advanced disease treated by the same modality. For radiation, there was no difference in all three functional categories for T1-2 vs. T3-4 (p = .84), showing that quality of life scores remain high for all stages. For surgery, functional status deteriorated significantly when comparing T1-2 vs. T3-4 (p = .0014), consistent with the fact that larger tumors require more extensive operations. CONCLUSION Radiation therapy provides a better performance status than surgery for base of tongue cancer. This is true for both early and advanced disease. Because radiation also provides similar local control and survival, our data suggests that radiation may be the preferred strategy. Functional scores remain high for all T stages treated with radiation, but deteriorate with more advanced T stages for patients treated with surgery. Similar studies using objective criteria are needed to further compare these treatments.

A prospective randomized trial of adjuvant brachytherapy in the management of low-grade soft tissue sarcomas of the extremity and superficial trunk.

PURPOSE This study was designed to evaluate the impact of adjuvant brachytherapy (BRT) on local and systemic recurrence rates in patients with low-grade sarcoma. PATIENTS AND METHODS Forty-five patients with histologic low-grade, completely resected soft tissue sarcomas of the extremity or superficial trunk were entered onto this trial. Following resection of all gross disease, patients were randomized to the BRT arm (n = 22) or to the no-BRT arm (n = 23). On the fifth or sixth postoperative day, catheters were loaded with iridium 192 to deliver a dose of 45 Gy to the tumor bed over 4 to 6 days. RESULTS The two groups were evenly distributed with respect to the distribution of presentation status (primary v recurrent), tumor site (trunk v extremity, proximal v distal extremity), tumor size (< 5 cm v > or = 5 cm), tumor depth (superficial v deep), and microscopic tumor margins (positive v negative). The predominant histopathologic diagnosis in each group was liposarcoma (BRT, 13 of 22 [59%]; no BRT, 14 of 23 [61%]) with other histopathologic subtypes evenly distributed between the two groups. The median follow-up duration among the ongoing survivors is 67 months. One patient in the BRT group developed systemic disease and died of progressive disease. Local recurrence occurred in five of 23 patients (22%) in the no-BRT group and six of 22 patients (27%) in the BRT group (P = .60). CONCLUSION Adjuvant radiation in the form of BRT does not appear to decrease local recurrence rates following complete resection of low-grade extremity and superficial trunk soft tissue sarcomas. Other adjuvant approaches, such as external-beam radiotherapy, are required to have a significant impact on local recurrence rates in this group of patients.

A prospective randomized trial of adjuvant chemotherapy with bolus versus continuous infusion of doxorubicin in patients with high‐grade extremity soft tissue sarcoma and an analysis of prognostic factors

A prospective randomized trial was conducted to compare the cardiotoxic and therapeutic effects of doxorubicin (60 mg/m2 every 3 to 4 weeks) administered by bolus or 72‐hour continuous infusion as adjuvant chemotherapy in 82 eligible patients after resection of high‐grade soft tissue sarcoma of the extremity or superficial trunk. Cardiac toxicity, defined as a 10% or greater decrease in left ventricular ejection fraction as assessed by radionuclide cineangiography, was evaluated in 69 patients. Cardiotoxicity was seen in 61% of patients in the bolus treatment arm with the median doxorubicin dose of 420 mg/m2. Among patients who received continuous infusion, 42% had cardiotoxicity with a median dose of 540 mg/m2. The rate of cardiotoxicity as a function of the cumulative dose of doxorubicin was significantly higher in the bolus treatment arm (P = 0.0017). Two patients in each group had clinical congestive heart failure, with one cardiac death occurring in each. There was a trend toward a lower rate of metastasis (P = 0.19) and a significantly lower rate of death of disease (P = 0.036) for patients treated with the bolus dose. Cox model analysis identified three unfavorable characteristics for the rate of developing a distant metastasis: blood transfusion within 24 hours of operation (P < 0.00001), tumor deep to the fascia and 5 cm or more in size (P = 0.0043), and a histologic subtype other than liposarcoma (P = 0.0002). The unfavorable effect of continuous infusion was not selected in the model (P = 0.16). Adjuvant chemotherapy for patients with soft tissue sarcoma is investigational. Furthermore, the impact of perioperative blood transfusion merits further study.

Prognostic factors in adenocarcinoma of the uterine cervix

A group of 136 female patients with adenocarcinoma of the uterine cervix, treated at Memorial Sloan‐Kettering Cancer Center between 1949 and 1981, with slides available for review, formed the basis for this study. They ranged in age from 10 to 91 years. Most (73%) had abnormal bleeding, either alone or in combination with other symptoms; 15% were asymptomatic. Eighty percent had a visible abnormality, most commonly an exophytic mass. Clinical stages were: 0 (3%), IB (61%), IIA (14%), IIB (10%), III (4%). There were four major histologic subtypes: mucinous (47%), endometrioid (24%), adenosquamous (15%), and clear cell (9%) carcinoma. Of the many clinicopathologic variables evaluated for prognosis, the most significant was stage of disease (P< 0.0001). Those with Stage IB disease had a survival probability of 76% at 5 years compared with 49% for those with Stage IIA and 34% with Stage IIB. The endometrioid pattern was associated with a more favorable prognosis than any other histologic subtype (P = 0.02). The presence of lymphatic tumor emboli and/or metastatic carcinoma in any lymph node group was associated with a less favorable prognosis for patients with Stage IB and IIA disease (P < 0.0001). Cancer 57:1584–1593, 1986.

Impact of adjuvant radiation on the patterns and rate of relapse in advanced-stage Hodgkin's disease treated with alternating chemotherapy combinations.

The role of adjuvant radiation therapy (RT) in the management of advanced-stage Hodgkins disease (HD) was analyzed in 222 patients who attained a complete remission (CR) with alternating chemotherapy combinations. Mechlorethamine, vincristine, procarbazine, and prednisone/doxorubicin, bleomycin, vinblastine, and dacarbazine (MOPP/ABVD) or MOPP/ABV alternating with the lomustine, melphalan, and vindesine combination (MOPP/ABV/CAD) were similarly effective in inducing a CR in 222 of 270 (83%) patients. These patients were scheduled to receive consolidative RT to bulky disease or other critical sites of initial nodal involvement to a total dose of 2,000 cGy, with an optional additional boost of 1,000 cGy. However, only 125 (56%) patients received radiation to all initial nodal sites of disease. In 69 (31%) patients, only selected nodal sites were included in the radiation fields, and 28 (13%) did not receive any RT. Of the 222 CR patients, 42 (19%) relapsed during a median follow-up period of 6.5 years (range, 2 to 15 years). Of these, 26 (62%) patients relapsed exclusively in unirradiated nodal sites, six (14%) within irradiated sites, and 10 (24%) both within and outside irradiated fields. The actuarial 10-year relapse-free survival (RFS) and overall survival (OS) for patients receiving radiation to all initially involved nodal sites were 89% and 94%, respectively, compared with 68% and 71% (P less than .0001) for patients who had only partial or no RT. Cox proportional hazards regression analysis showed that RT to all sites of initial disease was the most significant independent covariate (P less than .005) affecting RFS and OS. These data demonstrate that residual microscopic disease is relatively frequent in patients with apparent CR after alternating combination chemotherapy, and that irradiation of all sites of initial nodal involvement decreases relapse and improves survival in advanced-stage HD.

Preoperative and postoperative adjuvant combination chemotherapy for adults with high grade soft tissue sarcoma.

Background. Patients with high grade soft tissue sarcoma greater than or equal to 10 cm have a 3‐year disease‐free survival of approximately 30%. There is no convincing evidence, however, that postoperative adjuvant chemotherapy is beneficial. Preoperative chemotherapy has theoretical advantages over postoperative chemotherapy.