Ovarian cancer is the leading cause of death from a gynecologic malignancy in the United States. Most patients present with advanced disease and are treated with a combination of surgery and chemotherapy. Recently,
Ovarian cancer is the leading cause of death from a gynecologic malignancy in the United States. Most patients present with advanced disease and are treated with a combination of surgery and chemotherapy. Recently, paclitaxel has been identified as a highly active agent in this disease, and has produced a meaningful impact upon survival for patients with advanced disease. However, numerous questions remain regarding the optimal combination chemotherapy regimen and the integration of surgery into overall management. Furthermore, while patients with early-stage disease have an improved prognosis, there currently is no evidence that screening the general population is beneficial. Similarly, there is no evidence that any form of additional therapy for patients with advanced disease following their surgery and chemotherapy is beneficial. Controversy exists with regard to the benefit of second-look procedures to assess the impact of therapy and of close surveillance to select patients who may be candidates for additional therapy.
Ovarian cancer is a complex disease that requires knowledge of its biology and clinical course in order to select necessary diagnostic tests and to embark upon an appropriate treatment course. It is a disease of the elderly population and generally produces no specific signs or symptoms. Most patients are in their mid 60s when they are diagnosed, after a prolonged period of having vague abdominal complaints. Patients frequently undergo extensive radiologic studies unnecessarily, since a definitive diagnosis usually requires surgery, most frequently a laparotomy . A careful laparotomy is mandatory to stage the disease and to remove as much disease as possible.
At the time of diagnosis, most patients already have advanced-stage disease such that the disease has spread throughout the peritoneal cavity to involve multiple sites on peritoneal surfaces. About 75% of all patients will present with stage III disease. Stage IV disease, in which there is spread to the pleural or extraperitoneal sites such as the liver, has an even worse prognosis and, like stage III disease, is not curable by surgery. Subsequently, the vast majority of patients with ovarian cancer are treated immediately with chemotherapy following their initial surgery. Appropriate chemotherapy for ovarian cancer remains controversial. However, it appears that paclitaxel (Taxol) has made a significant impact upon the survival of patients with advanced disease, and the investigators in the Gynecologic Oncology Group (GOG) have now adopted the combination of paclitaxel plus a platinum compound as their standard therapy for ovarian cancer patients.
Such a combined modality approach using surgery and chemotherapy produces an approximately 70% to 85% response rate in patients with advanced disease . Unfortunately, the vast majority of these patients will ultimately recur, and 5-year survival rates for patients with advanced disease are somewhere around 25% to 30%. The appropriate treatment of patients at the time of recurrence and the follow-up of patients who do undergo a complete remission are also areas of controversy.
Because ovarian cancer has a markedly superior survival rate in patients who present with early-stage disease, a major effort has been underway to diagnose the disease in asymptomatic women with appropriate screening tests. However, this too has led to controversy: Some investigators are strong advocates of screening while others are concerned about its cost and potential deleterious effects.
Critical quality of life studies and outcomes analyses have not been frequently performed in ovarian cancer patients. This review points out those areas where such studies are needed. Clinical pathways need to be developed that give the patient the greatest opportunity for long-term survival and, at the same time, produce acceptable morbidity. Furthermore, outcomes analyses must be linked with economic studies to determine the cost effectiveness of practices and procedures used in the management of ovarian cancer patients.
This review examines current approaches to diagnosis, screening, initial treatment, follow-up, and management of recurrent disease, to focus on those areas that require further outcomes analyses and quality of life studies.
As noted, most patients with ovarian cancer are elderly with a protracted history of vague abdominal complaints. Unfortunately, many of these patients do not undergo pelvic examination at the time of their first visit to a physician or caregiver. When the patient does present with an adnexal mass, an upper abdominal mass, ascites, and/or pleural effusion, often an expensive diagnostic work-up is begun. Patients frequently undergo CT scans, ultrasound examinations, cystoscopies, barium enemas, and colonoscopies . The justification often given is to help the surgeon define the extent of normal tissue invasion prior to debulking surgery. Since most of these patients will go to surgery, it is unclear exactly how these preoperative studies are routinely beneficial. In addition, if there is any question about the diagnosis, a CA 125 test done in the presence of a pelvic mass in a postmenopausal woman has a high degree of specificity in correctly diagnosing epithelial ovarian cancer [3,4].
Most physicians would, at a minimum, require a chest x-ray and a CT scan of the abdomen prior to surgery, and reserve barium studies and in-depth examinations of the bowel for those patients in whom there is suspected involvement of the gastrointestinal tract.
Since patients who are diagnosed with early-stage ovarian cancer (ie, the disease has not spread beyond the pelvis) have a markedly higher survival rate, screening has been proposed to decrease morbidity and mortality from this disease [5,6]. The current screening tests being studied include transvaginal ultrasound, serum CA 125, and pelvic palpation. It should be emphasized that screening does not diagnose ovarian cancer. It only identifies patients who are at potential high risk for the disease, in whom a surgical procedure such as laparotomy is required to determine if they indeed have ovarian cancer.
The primary requirement for any screening procedure for ovarian cancer is that it have a high specificity, to avoid an excess of false-positive results leading to numerous unnecessary laparotomies. Although perhaps somewhat arbitrary, a minimum criterion for an acceptable screening procedure is a positive predictive value of 10% . (Positive predictive value is the ratio of true positives to true positives plus false negatives.) Even such a low positive predictive value means that there will be nine false-positive laparotomies for each diagnosed case of ovarian cancer.
Several large screening studies have been done [8-10], and these have recently been reviewed at the National Institutes of Health Consensus Conference on ovarian cancer . Due to the inadequate sensitivity and specificity of available screening tests, the current recommendation is that screening not be employed for the general population. This decision reflects three major factors:
1. The high rate of false-positive test results would lead to unnecessary operations and has the potential for doing harm in the screened population . Among patients undergoing unnecessary laparotomies, some would experience complications, and some, however small the number, might die from an unnecessary procedure. These deaths would likely outweigh any possible benefit from identifying a few patients with early-stage ovarian cancer in the screening protocol.
2. An elevated CA 125 level induces a great deal of anxiety in the patient and consequently leads to an unnecessary and expensive diagnostic workup, which often results in a negative laparotomy.
3. The cost of screening has been estimated to be about $15 billion if applied to the general US population . Consequently, while there may be a role for screening in patients with hereditary familial syndrome, women who are desirous of screening should participate in the ongoing National Cancer Institute PLCO trial in which 75,000 women will be randomized to screening with ovarian palpation, transvaginal ultrasound, and serum CA 125 or to routine follow-up. The study began in 1993, and initial results will be presented by the end of 1995. Dr. John Gohagan is coordinating the trial at the NCI.
The accepted surgical approach when the diagnosis of ovarian cancer is expected is a comprehensive laparotomy with a large midline incision that extends above the umbilicus. This will allow the gynecologic oncologist to adequately stage the cancer and at the same time perform cytoreduction. The standard operation includes a bilateral salpingo-oophorectomy and omentectomy as well as removal of any masses in the upper abdomen. The goal of cytoreductive surgery is to leave no tumor mass greater than 1 cm in diameter; if this is achieved, it is termed an optimal cytoreductive procedure.
Approximately 50% to 60% of all patients with advanced disease can undergo effective cytoreduction. However, controversy remains regarding the impact of cytoreductive surgery on the overall survival of these patients [14,15]. Nevertheless, it is accepted practice, since patients who undergo successful cytoreduction have a subsequently improved quality of life with decreased symptoms from large masses, and also respond more favorably to postoperative chemotherapy.
There has been substantial recent progress made in defining more effective chemotherapy for advanced ovarian cancer. Until recently, the combination of cyclophosphamide (Cytoxan, Neosar) and a platinum compound was considered standard care. This type of combination produced a response rate of approximately 55% to 60% in patients with suboptimal stage III and IV disease, and in this group of patients, median survival usually was 21 to 22 months. Carboplatin (Paraplatin) is a less toxic analog of cisplatin (Platinol), and many investigators have used carboplatin instead of cisplatin in combination with cyclophosphamide. A series of randomized trials and a metaanalysis have failed to demonstrate any significant difference in survival between patients treated with cisplatin or carboplatin [16-18]. Patients treated with carboplatin have a better quality of life due to decreased nausea and vomiting, less neurotoxicity, and less anorexia.
The taxanes represent a new family of antineoplastic agents that have altered the standard approach to ovarian cancer. The greatest clinical experience available is with paclitaxel (Taxol). This drug was initially derived from bark of the Western yew (Taxus brevifolia). Interest in this compound was heightened by its marked in vitro and in vivo activity in preclinical models of human cancer and by its novel mechanism of cytotoxicity . In contrast to other drugs that interfere with tubulin, paclitaxel leads to a polymerization of the microtubules, thereby disrupting mitoses.
Initial phase I trials with this agent were hampered by hypersensitivity reactions. Subsequently, by increasing the length of infusion to 24 hours and by premedicating the patients with steroids, diphenhydramine, and cimetidine (Tagamet), hypersensitivity reactions became infrequent .
In early phase II trials, paclitaxel was shown to induce a high response rate (approximately 30% to 40%) in previously untreated patients with advanced ovarian cancer [21,22]. Of particular note in these trials was the finding that even cisplatin-resistant patients were responding to paclitaxel. A subsequent phase I trial demonstrated that paclitaxel could be combined with cisplatin at full doses, and this led to a prospective randomized trial in untreated patients .
The GOG protocol 111 is a pivotal study in ovarian cancer in which patients with suboptimal stage III and IV disease were randomized to receive either six cycles of standard cisplatin plus cyclophosphamide or six cycles of cisplatin plus paclitaxel . After the completion of chemotherapy, patients were reassessed clinically, and those patients in a clinical complete remission underwent a second-look procedure to surgically assess their response to therapy. The patients randomized to the paclitaxel combination had a higher overall response rate (77% vs 62%), a higher clinical complete remission rate (54% vs 33%), a higher negative second-look or microscopic positive second-look rate (26% vs 41%), a longer time to disease progression (18 months vs 14 months), and, most important, a marked prolongation in median survival (37 months vs approximately 23 months). On the basis of these results, the GOG has accepted paclitaxel plus cisplatin as the standard chemotherapy regimen.
While this trial was being done, a subsequent clinical trial was evaluating the combination of paclitaxel plus carboplatin . This study was designed to establish a regimen with potentially less toxicity than cisplatin plus paclitaxel. Investigators at the Fox Chase Cancer Center performed a series of dose escalations with carboplatin and paclitaxel and have selected the following regimen for further evaluation in phase II and III trials: paclitaxel 175 mg/m² by a 3-hour infusion together with carboplatin dosed to an area under the curve (AUC) of 7.5. Since carboplatin excretion is dependent upon renal function, AUC dosing was developed to individualize a patient's carboplatin dose based upon her glomerular filtration rate (GFR) (or creatinine clearance). The dose of carboplatin (mg) is equal to the target AUC times the sum of the GFR (or creatinine clearance) plus 25, ie, dose = AUC[GFR + 25] . These cycles are administered every 21 days, and G-CSF (filgrastim, Neupogen) is not immediately required.
This study was associated with a very low incidence of severe hematologic toxicity, with platelet transfusions required in only 2% to 3% of cycles and hospitalizations for neutropenia or fever occurring in only 2% to 3% of cycles. Carboplatin plus paclitaxel had a very high response rate, with over 85% of patients responding, including approximately a 55% clinical complete response rate. On the basis of this study, the GOG will be incorporating this regimen into several new clinical trials.
Another important factor regarding paclitaxel that remains to be determined is the effect of dose intensity and administration schedule. Investigators at the NCI suggested that doses of 250 mg/m², which require G-CSF, will produce a higher overall response rate (48%) in previously treated patients . Furthermore, on the basis of preclinical data showing increased tumor kill when cells were exposed to paclitaxel for 96 hours, the NCI investigators performed a trial in doxorubicin-resistant breast cancer patients and reported a 48% response rate with 96-hour infusion of single-agent paclitaxel . Subsequently, Johnson et al at the NCI performed a phase I trial demonstrating that paclitaxel could be combined in a 96-hour infusion with cisplatin at a dose of 75 mg/m² (B. Johnson, personal communication).
Obviously, these approaches with paclitaxel, which require hematologic support and/or prolonged infusion, are more expensive and can have a negative impact on the quality of life of patients undergoing therapy, and clinical trials are needed to compare their efficacy with that of more standard approaches.
Based on these considerations, the GOG, as noted, has accepted paclitaxel plus a platinum compound as standard therapy. While many questions remain concerning how best to utilize paclitaxel, the impact upon survival observed in poor prognosis, suboptimal patients predicts that paclitaxel may have an even greater survival impact in the more favorable, optimal stage III patients.
The Table summarizes the ongoing GOG clinical trials for ovarian cancer. Quality of life studies and outcomes analyses will be incorporated into some of these trials.
This remains one of the most controversial areas in the management of patients with ovarian cancer. Approximately 60% to 70% of patients now will have no evidence of disease after cytoreductive surgery and six cycles of paclitaxel plus platinum-based chemotherapy. Unfortunately, many of these patients will recur. Numerous clinical trials are in progress to determine if any form of therapy can prevent or delay recurrences. Among the options being tested are intraperitoneal approaches with either chemotherapeutic agents, phosphorus 32 (32P), or immunotherapy. Alternatively, investigators are studying prolonged use of systemic non-cross-resistant drugs.
It has also been suggested that this group of patients be treated with high-dose chemotherapy, which requires hematologic support. Numerous clinical trials have been done in the United States, and well over 200 patients have now been treated with high-dose chemotherapy . These studies demonstrate that patients can respond to treatment with high drug doses, but that the duration of remission is relatively short. Currently, there is no evidence that this approach should be employed outside of clinical trials.
The appropriate follow-up of patients with advanced ovarian cancer after they achieve a clinical complete remission also remains controversial. Many investigators have performed second-look operations to surgically assess the completeness of the response. However, there is no evidence that this diagnostic procedure impacts upon the overall survival of patients with advanced ovarian cancer [30,31]. Similarly, the role of close surveillance in patients who achieve a clinical complete remission has been questioned. In many centers, it is current practice to obtain monthly serum CA 125 tests and CT scans of the pelvis and abdomen at regular intervals, as well as pelvic examinations. The major controversy regards the role of monitoring with serial CA 125 tests, particularly when the patient is asymptomatic, without radiologic evidence of disease, and starts to experience a rise in her CA 125. Should these patients be treated?
It has been suggested that these patients receive immediate "salvage" therapy. However, there is no true salvage therapy for patients who recur after initial chemotherapy, and virtually all of these patients are destined to die of their disease. Furthermore, patients who have a rising CA 125 may remain asymptomatic without evidence of radiologic regrowth of their tumors for many months and even, in some cases, for years. There is no evidence that treating patients when they are asymptomatic and only serologically positive is superior to waiting until their disease progresses to the point that it can be assessed radiologically or by physical examination.
While patients who do experience a long disease-free interval before recurrence can be successfully treated with drugs from their initial chemotherapeutic regimen, they ultimately will develop drug resistance. Clinical trials are in progress to understand the mechanisms of drug resistance and develop pharmacologic methods of reversal.
It has been reported that intraperitoneal scans using tagged antibodies can detect ovarian cancer deposits in the abdomen that cannot be detected by CT scans . Some have advocated that patients with a rising CA 125 should undergo such a procedure. However, this procedure has a significant false-positive rate, and even if one assumes that it correctly identifies a small focus that cannot be observed by CT scan, what is its clinical meaning? There is no convincing evidence that surgically removing the tumor, treating it with radiation, or starting immediate chemotherapy will impact upon the long-term survival of this patient. Consequently, it is our policy not to use such tests to follow patients in a clinical complete remission.
As noted, high-dose chemotherapy with autologous bone marrow transplantation or peripheral blood stem cell transfusions is an experimental approach. The use of peripheral blood stem cell transfusions allows administration of multiple cycles of chemotherapy, and preliminary reports from Memorial Sloan-Kettering, Fox Chase Cancer Center, and the University of North Carolina suggest that substantial response rates can be achieved in patients with advanced disease. However, this type of approach needs to be studied prospectively in a randomized trial, compared to standard therapy, in patients with small-volume ovarian cancer. Ovarian cancer patients should be encouraged to enter other experimental clinical trials as well. Phase I trials of new experimental drugs, as well as antibodies and antibody conjugates, are in progress. Such studies are important to help identify potentially novel therapeutic approaches.
This unfortunately remains an area of high priority in the overall management of patients with ovarian cancer. Most patients will not be cured of their disease, and, consequently, palliative care to maintain quality of life in a patient with little if any chance of a cure is an important medical consideration. Patients frequently develop bowel obstructions, which often raises the question of whether prolonged hyperalimentation should be instituted. A secondary debulking in a patient who has failed chemotherapy is clearly ineffective. However, loop colostomies and bypass procedures can be important in maintaining quality of life for these patients. In addition, judicious use of radiation therapy for symptomatic metastases has been shown to provide effective palliation . Similarly, narcotics are frequently necessary in the terminal stages of the disease, and appropriate pain control cannot be overemphasized as a critical factor in maintaining quality of life.
The cure rates for ovarian cancer are slowly improving, particularly with the use of paclitaxel. It must be emphasized that there is only one chance to cure patients, and that is with the initial therapeutic approach. Consequently, all patients with advanced ovarian cancer who require chemotherapy should be treated with a combination of paclitaxel plus a platinum compound. Patients should be encouraged to enter clinical trials that are exploring the optimum platinum compound to use as well as the best dose and schedule of paclitaxel. At the completion of induction chemotherapy, patients should be informed about the risk of recurrence and the controversy regarding close surveillance.
A frank discussion regarding the role of second-look operations is also mandatory. There is a true difference of opinion among physicians regarding the role of second-look laparotomy and its potential benefit. The NIH Consensus Conference failed to endorse the routine use of second-look laparotomies and suggested they be reserved primarily for patients on clinical trials. When patients suffer a disease recurrence, they should be encouraged to enter appropriate clinical trials addressing whether any of the experimental treatments, such as intraperitoneal therapy and high-dose chemotherapy, may be beneficial.
1. Brooks SE: Preoperative evaluation of patients with suspected ovarian cancer. Gynecol Oncol 55:S80-S90, 1994.
2. Ozols RF: Ovarian cancer. Part II: Treatment. Current Probl Cancer 16:65-126, 1992.
3. Finkler NJ, Benacerraf B, Lavin PT, et al: Comparison of serum CA 125, clinical impression, and ultrasound in the preoperative evaluation of ovarian masses. Obstet Gynecol 72:659-671, 1988.
4. Malkasian HGD, Knapp RC, Lavin PT, et al: Preoperative evaluation of serum CA 125 levels in premenopausal and postmenopausal patients with pelvic masses: Discrimination of benign from malignant disease. Am J Obstet Gynecol 159:341-346, 1988.
5. DePriest PD, van Nagell JR, Gallion HH, et al: Ovarian cancer screening in asymptomatic postmenopausal women. Gynecol Oncol 51:205-209, 1993.
6. DePriest PD, Shenson BS, Fried A, et al: A morphology index based on sonographic findings in ovarian cancer. Gynecol Oncol 51:7-11, 1993.
7. Jacobs I: Genetic, biochemical, and multimodal approaches to screening for ovarian cancer. Gynecol Oncol 55:S22-S27, 1994.
8. Jacobs IJ, Stabile I, Bridges J, et al: Multimodal approach to screening for ovarian cancer. Lancet 1:268-271, 1988.
9. Jacobs I, Davies AP, Bridges J, et al: Prevalence screening for ovarian cancer in postmenopausal women by CA 125 measurement and ultrasonography. Br Med J 306:1030, 1993.
10. Einhorn N, Sjovall K, Knapp RC, et al: Prospective evaluation of serum CA 125 levels for early detection of ovarian cancer. Obstet Gynecol 80:14-18, 1992.
11. Kramer BS, Gohagan J, Prorok PC: NIH Consensus 1994: Screening. Gynecol Oncol 55:S20-S21, 1994.
12. Schapira MM, Matcher DB, Young MJ: The effectiveness of ovarian cancer screening: A decision analysis model. Ann Intern Med 118:838-843, 1993.
13. Creasman WJ, DiSala PJ: Screening in ovarian cancer. Am J Obstet Gynecol 165:7-10, 1991.
14. Hoskins WJ, Bundy BN, Thigpen JT, et al: The influence of cytoreductive surgery on recurrence-free interval and survival in small-volume stage III epithelial ovarian cancer: A Gynecologic Oncology Group study. Gynecol Oncol 47:159-166, 1992.
15. Hunter RW, Alexander NDE, Soutter WP: Meta-analysis of surgery in advanced ovarian carcinoma: Is maximum cytoreductive surgery an independent determinant of prognosis? Am J Obstet Gynecol 166:504-511, 1992.
16. Swenerton K, Jeffrey J, Stuart G, et al: Cisplatin-cyclophosphamide versus carboplatin-cyclophosphamide in advanced ovarian cancer: A randomized phase III study of the National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol 10:718-726, 1992.
17. Alberts DS, Green S, Hannigan EV, et al: Improved therapeutic index of carboplatin plus cyclophosphamide versus cisplatin plus cyclophosphamide: Final report by the Southwest Oncology Group of a phase III randomized trial in stages III and IV ovarian cancer. J Clin Oncol 10:706-717, 1992.
18. Advanced Ovarian Trialists Group: Chemotherapy in advanced ovarian cancer: An overview of randomised clinical trials. Br Med J 303:884-893, 1991.
19. Donehower RC, Rowinsky EK: Paclitaxel, in Devita VT, Rosenberg SA, Hellman S: Principles & Practice of Oncology. PPO Updates, vol. 8, no. 10, pp 1-16. Philadelphia, JB Lippincott, 1994.
20. Trimble EL, Adams JD, Vena D, et al: Paclitaxel for platinum-refractory ovarian cancer: Results from the first 1,000 patients registered to National Cancer Institute Treatment Referral Center 9103. J Clin Oncol 11:2405-2410, 1993.
21. Einzig AI, Wiernik PH, Sasloff J, et al: Phase II study and long-term follow-up of patients treated with Taxol for advanced ovarian adenocarcinoma. J Clin Oncol 10:1748-1753, 1992.
22. Thigpen JT, Blessing JA, Ball H, et al: Phase II trial of paclitaxel in patients with progressive ovarian carcinoma after platinum-based chemotherapy: A Gynecologic Oncology Group study. J Clin Oncol 12:1748-1753, 1994.
23. Rowinsky EK, Gilbert MR, McGuire WP, et al: Sequences of Taxol and cisplatin: A phase I and pharmacologic study. J Clin Oncol 9:1692-1703, 1991.
24. McGuire WP, Hoskins WJ, Brady MF, et al: A phase III trial comparing cisplatin/Cytoxan (PC) and cisplatin/Taxol (PT) in advanced ovarian cancer (AOC) (abstract). Proc Am Soc Clin Oncol 12:255, 1993.
25. Ozols RF, Kilpatrick D, O'Dwyer P, et al: Phase I and pharmacokinetic study of Taxol (T) and carboplatin (C) in previously untreated patients (PTS) with advanced epithelial ovarian cancer (OC): A pilot study of the Gynecology Oncology Group (abstract). Proc Am Soc Clin Oncol 12:259, 1993.
26. Calvert AH, Newell DR, Gumbrell LA, et al: Carboplatin dosage: Prospective evaluation of a simple formula based on renal function. J Clin Oncol 7:1748-1756, 1989.
27. Kohn EC, Sarosy G, Bicher A, et al: Dose-intense Taxol: High response rate in patients with platinum-resistant recurrent ovarian cancer. J Natl Cancer Inst 86:18-24, 1994.
28. Wilson WH, Berg SL, Bryant G, et al: Paclitaxel in doxorubicin-refractory or mitoxantrone-refractory breast cancer: A phase I/II trial of 96-hour infusion. J Clin Oncol 12:1621-1629, 1994.
29. Schilder RJ: High-dose chemotherapy with autologous hematopoietic cell support in gynecologic malignancies, in Hoskins WJ, Perez CA, Young RC: Principles and Practice of Gynecologic Oncology Updates, vol. 1, no. 3, pp 1-14 Philadelphia, JB Lippincott, 1993.
30. Friedman JB, Weiss NS: Second thoughts about second-look laparotomy in advanced ovarian cancer. N Engl J Med 322:1079-1081, 1990.
31. Miller DS, Spirtos NM, Ballon SC, et al: Critical reassessment of second-look exploratory laparotomy for epithelial ovarian carcinoma: Minimal diagnostic and therapeutic value in patients with persistent cancer. Cancer 69:502-510, 1992.
32. Krag DNJ: Clinical utility of immunoscintigraphy in managing ovarian cancer. J Nucl Med 34:545-548, 1993.
33. Corn BN, Lanciano R, Ozols RF, et al: Recurrent ovarian cancer: Effective palliation of chemotherapy failures by irradiation. Cancer 74:2979-2983, 1994.