Donald M. Stablein

Intracerebral metastases in solid‐tumor patients: Natural history and results of treatment

In order to determine the natural history and results of treatment of intracerebral metastases in solid‐tumor patients, the records of 191 patients with an antemortem diagnosis of intracerebral metastasis made during the period from August 1974 to November 1978 were reviewed. Malignancies included lung (122 patients), breast (26), unknown primary (16), melanoma (8), colorectal (6), hypernephroma (4), and others (12). Favorable prognostic factors included solitary brain metastasis (P < 0.001), ambulatory performance status (P < 0.001), symptoms of headache (P < 0.001), or visual disturbances (P < 0.02), and estrogen receptor positivity in breast cancer patients (P = 0.055). Poor prognostic factors included advanced age (P < 0.04) and evidence of impaired consciousness, i.e., disorientation, lethargy, stupor, or coma (P < 0.007). Median survival time after diagnosis of intracerebral metastasis was 3.7 months for the entire series. In those patients with a single intracerebral metastasis and minimal tumor burden, the type of treatment used had a significant impact on survival. Those cases treated with surgery and radiation had a median survival time of 9.7 months versus 3.7 months for those treated with radiation alone (P < 0.02). When using a proportional hazard regression analysis to adjust for the three most important prognostic factors, treatment (surgery and radiation versus radiation alone) still appeared to be important. Intracerebral metastases were the immediate or contributing cause of death in 50% of the patients in this series. Patients at greater risk of dying of intracerebral metastases included those in whom the brain was the first site of distant metastasis, those with an intracerebral metastasis from an unknown primary site, and those whose presentation of malignancy was with symptoms of a brain metastasis. Although the therapeutic goal in intracerebral metastases is generally palliative, it appears that there are categories of cases that may benefit from more aggressive treatment.

Statistical Methods for Determining Prognosis in Severe Head Injury

Determining the prognostic significance of clinical factors for patients with severe head injury can lead to an improved understanding of the pathophysiology of head injury and to improvement in therapy. A technique known as the sequential Bayes method has been used previously for the purpose of prognosis. The application of this method assumes that prognostic factors are statistically independent. It is now known that they are not. Violation of the assumption of independence may produce errors in determining prognosis. As an alternative technique for predicting the outcome of patients with severe head injury, a logistic regression model is proposed. A preliminary evaluation of the method using data from 115 patients with head injury shows the feasibility of using early data to predict outcome accurately and of being able to rank input variables in order of their prognostc significance.

Analysis of survival data with nonproportional hazard functions.

The log-rank test or the proportional hazard model is a valuable, widely accepted method for analyzing time-to-response data from comparative clinical trials. When the hazard ratio is constant in time, this procedure is optimal. Indiscriminate or unthinking use of this approach results in problems in the determination of treatment differences. For example, when the true survival curves intersect, the hazard ratio cannot be constant, i.e., the hazard functions are not proportional. It is shown that by considering time-by-treatment interactions we gain flexibility in describing the relationships among hazard functions. In this paper we demonstrate with the results of a clinical trial how available methodology can be used to permit tests for the appropriateness of the model and to enable informative analysis of such data.

A two-sample test sensitive to crossing hazards in uncensored and singly censored data.

Savage score statistics are employed to develop a test for comparing survival distributions with right-hand singly censored data. The procedure is motivated by the interest in developing a powerful method for determining differences when true survival distributions cross. Examination of small-sample characteristics under the null hypothesis indicate that asymptotic critical values yield a slightly conservative test. Power of the test compares favorably with other criteria, including the modified Smirnov procedure, particularly if there is a single crossing of the survival curves.

Confidence regions for constrained optima in response-surface experiments.

The precision of the estimated optimum from a response-surface experiment is often indicated via a confidence region about the optimum. Sometimes, because of associated secondary responses, unconstrained optima produce unrealistic operating conditions, even when the true response surface is known. We consider confidence intervals for constrained optima for which the constraint function is known or separately estimated. An example from a cancer combination chemotherapy experiment illustrates the construction of such a region.

Hyperfractionated radiation and chemotherapy for unresectable localized adenocarcinoma of the pancreas: the gastrointestinal tumor study group experience

Eighteen patients with unresectable localized adenocarcinoma of the pancreas were treated by a combination of chemotherapy plus hyperfractionated radiation therapy to the pancreas for 4080 cGy with an additional 960 cGy to the pancreatic tumor and a surrounding margin. One hundred and twenty cGy were given twice daily 4 to 6 hours apart. High‐energy photon or electron beams were used with treatment planning based on computed tomographic (CT) scans. Patients were given chemotherapy in the form of 5‐fluorouracil (5‐FU) at 350 mg/m2 on the first 3 and last 3 days of radiation therapy. On day 53, chemotherapy was given that included 600 mg/m2 IV of 5‐FU, 1 gm/m2 of streptozotocin, and 10 mg/m2 IV of mitomycin C. The 5‐FU and streptozotocin were repeated on days 60, 81, and 88, and the stretozotocin and mitomycin (SMF) cycles were repeated every 8 weeks until progression. Radiation toxicity was generally tolerable with one of 18 evaluable patients having severe nausea and vomiting and two of 18 patients having severe diarrhea. One patient had total liver failure and died 3 months after initiation of therapy. Six patients had severe hematopoietic toxicity during chemotherapy. Overall, the severe toxicity rate was higher (67%) than in previous studies. Median survival was 35 weeks, the 1‐year survival rate was 39%, and the patient who survived the longest died at 68 months. Although this schedule of hyperfractionated radiation and chemotherapy was disappointing, combined experimental radiation approaches plus chemotherapy for localized unresectable adenocarcinoma of the pancreas deserve additional research.

The 1987 USA National CAPD Registry Report

The National CAPD Registry has noted encouraging trends in CAPD therapy in the U.S. Recently, trained patients have reported lower peritonitis rates, fewer days hospitalized for dialysis-related events, and lower probability of developing a first episode of a complication. If accumulating experience, better patient selection, and new technologic developments further these trends, then the relatively high transfer rates should begin to decrease since many transfers are related to peritonitis and other complications. The Registrys observations suggest that CAPD is an evolving therapy and that as complication rates and transfer rates decrease, CAPD may play an even greater role in the treatment of patients with end-stage renal disease.

Survival analysis of drug combinations using a hazards model with time-dependent covariates.

Hazard functions in cancer chemotherapeutic situations may not be proportional, so a nonproportional hazard model has been developed. The dose-response surface is explored by regression analysis of experimental data, and after the estimation of the underlying hazard function the quality of the fit of the model is assessed. Further, treatment levels may be optimized, and estimated survival distributions can be plotted for any treatment combination. In an example of two-drug treatment of murine L1210 leukemia, statistically significant nonproportionality is determined. Analysis permits extraction of potentially important information on drug interrelationships, which has been previously unavailable.

Timing and characteristics of multiple peritonitis episodes: a report of the National CAPD Registry.

Patterns of recurrent peritonitis episodes were examined in 6,335 new continuous ambulatory peritoneal dialysis (CAPD) patients entered into the National CAPD Registry. Forty-six percent of all peritonitis episodes were initial occurrences, with 8% of the patients reporting four or more episodes. The proportion of gram-positive and gram-negative infections was constant across episodes. In patients with multiple infections, negative organisms were found to have increased risk of recurring as gram-negative infection. A similar observation was made for fungal infections. Of patients with multiple peritonitis episodes, more than 40% of those who transferred to other maintenance renal replacement therapy identified peritonitis as the reason for transfer. A discrete time logistic model was used to estimate peritonitis risk in 4-month follow-up periods. Patients like those on the registry are estimated to have a 22% risk of developing peritonitis during any 4-month period. This risk was increased 4% for patients aged less than 21 years, 7% for nonwhite patients, and 19% in the period following a peritoneal infection.

Perineal effects of postoperative treatment for adenocarcinoma of the rectum

Nine (4%) first recurrences that involved the perineum were identified in a randomized study of 202 patients treated by no further therapy, chemotherapy only, radiotherapy only, combined radiotherapy and chemotherapy, following complete surgical excision of adenocarcinoma of the rectum. Six of these were in unirradiated patients and in two of the three irradiated patients the perineum was included in the treatment volume. Eight of the nine patients were male and all nine had received abdominoperineal resection (APR). Our quality assurance procedures identified 22 of 96 irradiated patients in whom the perineum was grossly outside the fields. Sixteen of these had undergone APR. As only one of these 16 relapsed in the perineum no definite effect of the surgical procedure on the likelihood of perineal recurrence could be demonstrated. Examination of the pathology reports revealed that 28 patients undergoing APR had tumors within 2 cm of the anorectal junction (pectinate line). Five (17.8%) developed perineal recurrence compared with 4 (3.6%) of 110 patients whose tumors were more than 2 cm from the anus (p less than 0.02-Fisher exact test). No survival differences could be demonstrated between those receiving perineal irradiation and those not but perineal irradiation was associated with toxicity with at least nine (12.2%) out of 74 developing severe complications directly related to the perineum. The routine inclusion of the perineum in postoperative pelvic irradiation fields for all cases of adenocarcinoma of the rectum is questioned. Our current policy following APR includes optional coverage of the perineum for those tumors more than 5 cm from the anorectal junction.