There is no role for hyperfractionated radiotherapy in the management of children with newly diagnosed diffuse intrinsic brainstem tumors: results of a Pediatric Oncology Group phase III trial comparing conventional vs. hyperfractionated radiotherapy

PURPOSE

In June 1992, POG began accrual to a phase III study, POG-9239, designed to compare the time to disease progression, overall survival, and toxicities observed in children with newly diagnosed brainstem tumor treated with 100 mg/m2 of infusional cisplatin and randomized to either conventional vs. hyperfractionated radiotherapy.

METHODS AND MATERIALS

Patients eligible for study were those between 3 and 21 years of age with previously untreated tumors arising in the pons. Histologic confirmation of diagnosis was not mandatory, provided that the clinical and MRI scan findings were typical for a diffusely infiltrating pontine lesion. Treatment consisted of a six-week course of local field radiotherapy with either once a day treatment of 180 cGy per fraction to a total dose of 5400 cGy (arm 1) or a twice a day regimen of 117 cGy per fraction to a total dose of 7020 cGy (the second of the three hyperfractionated dose escalation levels of POG-8495) (arm 2). Because of previously reported poor results with conventional radiotherapy alone, cisplatin was included as a potential radiosensitizer in an attempt to improve progression-free and ultimate survival rates. Based on results of the phase I cisplatin dose escalation trial, POG-9139, 100 mg/m2 was chosen for this trial and was delivered by continuous infusion over a 120-hour period, beginning on the first day of radiotherapy and repeated during weeks 3 and 5. One hundred thirty eligible patients were treated on protocol, 66 on arm 1 and 64 on arm 2.

RESULTS

The results we report are from time of diagnosis through October 1997. For patients treated on arm 1, the median time to disease progression (defined as time to off study) was 6 months (range 2-15 months) and the median time to death 8.5 months (range 3-24 months); survival at 1 year was 30.9% and at 2 years, 7.1%. For patients treated on arm 2, the corresponding values were 5 months (range 1-12 months) and 8 months (range 1-23 months), with 1- and 2-year survival rates at 27.0% and 6.7%, respectively. Evaluation of response by MRI at 4 or 8 wks post treatment was available in 108 patients and revealed a complete response in 1 patient of each Rx arm, a partial response (> 50% decrease in size) in 18 patients of arm 1 and 15 patients of arm 2, minimal to no response (stable) in 25 patients of arm 1 and 23 patients of arm 2, and progressive disease in 13 patients of arm 1 and 12 patients of arm 2. The pattern of failure was local in all patients. Morbidity of treatment was similar in both Rx arms, with no significant toxicity (including hearing loss) reported. Autopsy was performed in 6 patients, and confirmed the presence of extensive residual tumor in these cases.

CONCLUSION

The major conclusion from this trial is that the hyperfractionated method of Rx 2 did not improve event-free survival (p = 0.96) nor did it improve survival (p = 0.65) over that of the conventional fractionation regimen of Rx 1, and that both treatments are associated with a poor disease-free and survival outcome.

The role of thallium-201 uptake and retention in intracranial tumors after radiotherapy

This study prospectively assessed the diagnostic accuracy and prognostic value of 201TI uptake and retention in primary and metastatic intracranial tumors treated by conventional radiotherapy and/or radiosurgery.

METHODS

An initial 201TI study (early and delayed images), was obtained in 60 postsurgical patients, 6-12 wk after radiotherapy or radiosurgery. Repeat imaging was performed as clinically warranted. Tumor-to-background count ratios and a retention index (RI) were calculated for all lesions.

RESULTS

Abnormally increased 201TI uptake was observed in 40 of 60 patients. In all patients with positive results, the diagnosis of residual tumor was confirmed at biopsy or by clinical follow-up. In 20 of 60 patients, no abnormal 201TI uptake was observed, despite findings on CT and/or MRI scans that were suspicious for tumor. Ten of the negative 201TI studies were confirmed as true-negatives by the clinical course and by resolution of CT/MRI abnormalities. The remaining 10 negative SPECT studies ultimately proved to be false-negatives: six of these patients had lesions < 1 cm in maximum diameter, one patient had a large metastatic choriocarcinoma; and three patients had low-grade astrocytomas > 2 cm in minimum diameter. Tumor-to-background ratio of 201TI uptake did not distinguish between tumor type, or predict clinical outcome. The RI of 201TI was significantly higher for metastatic melanoma than for other tumor metastases. It demonstrated reasonably good correlation with clinical outcome: 6/7 patients with eventual tumor regression showed a decrease in RI on follow-up examination, and 4/5 patients with eventual tumor progression had an increase in RI.

CONCLUSION

Thallium-201 brain SPECT appears to be a useful noninvasive imaging technique in patients irradiated for intracranial tumors. Thallium-201 scintigraphy has very high specificity (100% in this cohort) for detecting viable residual tumor. False-negative findings may occur. Quantitative analysis of 201TI uptake has limited diagnostic and prognostic significance, but changes in 201TI retention after radiation therapy seems to have prognostic value.

Hyperfractionated irradiation and concurrent cisplatin in brain stem tumors: a Pediatric Oncology Group pilot study (9139)

A phase I Pediatric Oncology Group Study was performed combining 7,020 cGy hyperfractionated irradiation (117 cGy twice daily separated by 6 h) with increasing concurrent doses of cisplatin given with the intent of radiosensitization as treatment for children with newly diagnosed brain stem tumors. Cisplatin was infused over 120 h on weeks 1, 3, and 5 during a 6-week radiotherapy course. The following cisplatin dose levels were studied: (1) 50 mg/m2/120 h, (2) 75 mg/m2/120 h and (3) 100 mg/m2/120 h. Sixteen of 17 children completed therapy. One child expired after 2 days of treatment secondary to massive intratumoral hemorrhage. At cisplatin dose level 3 (100 mg/m2/120 h), grade 2-4 myelosuppression was encountered in 3 of 5 evaluable patients. Otherwise, no other excessive toxicities, including renal and ototoxicity, were noted.

Ongoing progress in the treatment of childhood rhabdomyosarcoma

During the past 2 decades, several different multidisciplinary treatment studies of childhood rhabdomyosarcoma have been conducted, some as single-institution protocols and others as multi-institution cooperative endeavors. Much of our current understanding of the natural history and histopathology, as well as response to treatment, is derived from these studies. This understanding has allowed for the continued evolution of treatment, which has already resulted in a dramatic increase in overall survival rates. Current treatment trials are aimed at maintaining high cure rates while reducing treatment-related morbidity, and exploring innovative regimens to improve the outlook of poor-risk patients. The improved outcome in children with rhabdomyosarcoma is a tribute to the multi-institutional, multidisciplinary approach to cancer research and treatment.

Myeloablative combination chemotherapy without total body irradiation for neuroblastoma

Myeloablative treatment intensification in 25 patients diagnosed when older than 12 months of age with stage IV neuroblastoma included sequential delivery of cisplatin 120 mg/m2 x 1, hyperfractionated radiation (2,100 cGy) to the primary site and adjacent lymph nodes, carmustine (BCNU) 200 mg/m2 x 1, melphalan 60 mg/m2/d x 3 (n = 13) or thiotepa 300 mg/m2/d x 3. (n = 12), and etoposide (VP 16) 300 mg/m2/d x 3. Seventy-two hours after the last dose of VP 16, histologically tumor-free and 4-hydroperoxycyclophosphamide (4-HC; 100 mumol/L)-purged autologous bone marrow (ABMT) was infused. Acute toxicities included grade 3 to 4 oral mucositis, grade 1 to 2 diarrhea, and fevers. No patient required infusion of unpurged reserve autografts. At ABMT, 16 patients (group I) were progression-free 6.5 months to 14 months (median, 9 months) from diagnosis: seven remain progression-free 20 months to 46 months (median, 39 months) off therapy, six relapsed 4 months to 17 months post-ABMT, and three died of toxicity (candidiasis, metabolic derangement, and venoocclusive disease [VOD]). The event-free survival of group I patients is 44% at 24 months post-ABMT. Nine patients (group II) were in second remission at ABMT, including three who had relapsed after other transplant procedures: two are progression-free 24 months and 41 months off therapy, four relapsed 3 months to 12 months post-ABMT, and three died of toxicity (aspergillosis, hemorrhagic cystitis, VOD). Only one of 10 relapses involved a primary site, suggesting a beneficial effect of local radiation. In terms of survival or toxicity, an advantage for melphalan or thiotepa was not evident. Regimens such as this may prolong the survival of selected patients with poor-risk neuroblastoma, but concerns over late relapses and toxicity mandate continuing efforts to devise alternative, less risky, and more clearly beneficial approaches for definitive ablation of neuroblastoma.

Delayed central nervous system (CNS) radiation in childhood CNS acute lymphoblastic leukemia. Results of a pilot trial

With the introduction of effective prophylactic central nervous system (CNS) therapy in childhood acute lymphoblastic leukemia (ALL), the incidence of CNS relapse has been significantly reduced. Nonetheless, there is a continuing need for effective CNS therapy for the 5% to 10% of children who escape prophylaxis and in whom active CNS disease develops. In July 1985, a pilot study was initiated whereby the consolidative CNS treatment, given in the form of craniospinal axis radiation, was delayed and administered after delivery of a prolonged course of intensive systemic and CNS chemotherapy. Ten leukemic patients with CNS relapse were treated according to this pilot study. One patient did not respond to pre-CNS consolidative therapy and died at 2 months with progressive disease. The remaining nine patients have completed the craniospinal axis radiation. Five patients are off all therapy and remain in remission at a median time of 38 months (range, 31 to 46 months). Two patients are near completion of therapy and are disease-free at 22 and 24 months. Testicular relapse occurred in two patients at 14 and 29 months. CNS chemotherapy, as used in this trial, appears to have allowed for a delay in the administration of the consolidative craniospinal radiation without negatively affecting the CNS remission rate or duration. In turn, this delay has allowed for the uncompromised delivery of intensive multiagent systemic chemotherapy, as well as the separation of those patients destined to early systemic relapse from those who will achieve a sustained complete remission. In both cases, a reduction in the need for multiple courses of potentially toxic CNS therapy as a result of CNS reseeding after radiation is anticipated.

Factors predictive of mortality in pediatric extremity rhabdomyosarcoma

In order to examine factors predictive of fatal outcome in children presenting with histologically confirmed extremity rhabdomyosarcoma, we performed a retrospective analysis of our institutional experience from 1970 to 1985. Thirty-five patients were identified and staged according to international criteria (TNM). Variables evaluated for their predictive effect on fatal outcome included (1) tumor invasiveness, (2) tumor size, (3) anatomic location of the primary, (4) regional lymph node involvement, (5) distant metastases at presentation, (6) complete surgical resection, (7) use of amputation, and (8) alveolar histologic subtype. Significant predictors of mortality included (1) tumor invasiveness (P less than or equal to .0001), (2) regional node involvement (P less than or equal to .0002), (3) distant metastases at the time of presentation (P less than or equal to .001), (4) alveolar histology (P less than or equal to .001), (5) size of primary (P less than or equal to .007), and (6) completeness of surgical resection (P less than or equal to .05). In multivariate analysis, local tumor invasiveness was the most important predictor of fatal outcome with an associated relative risk of 18. We conclude that local tumor invasiveness is the most important determinant of clinical stage.

Prognosis in childhood rhabdomyosarcoma of the extremity

From 1970 to 1987, 34 patients younger than 22 years of age with extremity rhabdomyosarcoma were treated at the Memorial Sloan-Kettering Cancer Center (MSKCC). All patients were treated according to protocols consisting of surgery, radiotherapy, and multiple drug chemotherapy. Sixteen of 34 patients survived, and 14 continue to be disease-free; the 5-year survival rate was 44%. A retrospective univariate analysis of data according to the TNM staging system was undertaken. Tumor invasion, regional nodal involvement, distant metastases, and alveolar histologic condition each had a significant impact on survival. However, in multivariate analysis, the stage of disease at diagnosis was the most important predictor of survival outcome.

The influence of extensive bone erosion on local control in non-orbital rhabdomyosarcoma of the head and neck

To investigate a recent report suggesting extensive bone erosion (EBE) as an important prognostic factor in head and neck rhabdomyosarcoma, a retrospective review was performed of 32 patients with gross residual and/or metastatic non-orbital head and neck rhabdomyosarcoma treated between the years 1971-1987. Treatment consisted of surgery, radiation therapy (median primary dose, 5000 cGy) and combination chemotherapy according to the Memorial Sloan-Kettering Cancer Center (MSKCC) T2 and T6 protocols. The MSKCC staging system was used: Stage II, 23 patients; Stage III, 6 patients; and Stage IV, 3 patients. With a median follow-up of 48 months (range 17 months to 13 years), the overall survival and local control rates were 72% (23/32) and 75% (24/32), respectively. Local control was achieved in 11/11 patients without extensive bone erosion (Stages II, 9/9; III, 1/1; IV, 1/1) as compared to 13/21 patients (Stages II, 10/14; III, 2/5; IV, 1/2) with extensive bone erosion (p = 0.02). Our data appear to support the recently reported finding that extensive bone erosion is an important predictor of local failure.

Mortality in pediatric paratesticular rhabdomyosarcoma: a multivariate analysis

We performed a retrospective analysis of 28 children presenting with confirmed paratesticular embryonal rhabdomyosarcoma to examine factors predictive of fatal outcome. Complete surgery, defined as radical inguinal orchiectomy with clear microscopic margins plus radical, retroperitoneal lymph node dissection with extirpation of all gross disease, was performed in 21 patients (75 per cent). All patients received chemotherapy and 20 received radiation therapy according to protocol. Of the patients 16 have survived for more than 5 years with no evidence of disease. Univariate analysis revealed that only the presence of parenchymal metastases at diagnosis (p less than or equal to 0.040) and unresectability (p less than or equal to 0.003) were significant predictors of fatal outcome. Multivariate analysis showed that unresectability was the most important predictor of mortality with an estimated relative risk of 5.8.

The role of postoperative radiotherapy after resection of single brain metastases

To assess the value of whole brain radiotherapy (WBRT) after complete resection of a single brain metastasis we reviewed the records of 98 patients who had elective craniotomy between 1978 and 1985. Seventy-nine patients received postoperative WBRT (Group A) and 19 patients no radiotherapy (RT) (Group B). Neurological relapse was designated as local (i.e., at the site of the original metastasis) or distant (i.e., elsewhere in the brain). Postoperative WBRT significantly prolonged the time to any neurological relapse (P = 0.034) with a 1-year recurrence rate of 22% in Group A and 46% in Group B patients; however, it did not specifically control either local or distant cerebral recurrence. Recurrence of metastatic brain disease was not affected by location of the original lesion; however, meningeal relapse occurred in 38% of cerebellar lesions, but only in 4.7% of supratentorial metastases (P = 0.003). The total radiation dose or fractionation scheme of RT did not affect survival nor time to neurological relapse. The median survival was 20.6 and 14.4 months for Groups A and B, respectively (not statistically different). Forty-eight percent of Group A and 47% of Group B patients survived for 1 year or longer; however, 11% of patients who had received RT and survived 1 year developed severe radiation-induced dementia. All patients with radiation-related cerebral damage received hypo-fractionated RT with high daily fractions as commonly designed for rapid palliation of macroscopic brain metastases.(ABSTRACT TRUNCATED AT 250 WORDS)

Reduced incidence of the somnolence syndrome in leukemic children with steroid coverage during prophylactic cranial radiation therapy. Results of a pilot study

Chemotherapeutic regimens for childhood acute lymphoblastic leukemia (ALL) include a remission induction period with high, daily doses of prednisone among other agents. A period of central nervous system (CNS) prophylaxis follows, during which steroids are often tapered entirely before cranial radiation (CRT) is completed or even initiated. The somnolence syndrome (SS) has been described 4 to 6 weeks after completion of CRT in up to 60% of the children with doses as low as 1800 cGy. A pilot study of continuous steroid coverage during CRT in childhood ALL was conducted. From July 1984 to July 1986, 38 children entered on Children's Cancer Study Group ALL protocols received CRT of 1800 cGy (180 cGy x 10). All patients received oral prednisone throughout the entire course of CRT at daily doses varying from 3.0 to 60.0 mg/m2. The overall incidence of the SS was 13% (five patients). The development of the syndrome was steroid dose-dependent: greater than or equal to 15 mg/m2/d (one of 32 patients), 3% incidence; less than 15 mg/m2 (four of six patients), 67% incidence. The presence of headache during CRT was also steroid dose-related: greater than or equal to 15 mg/m2, one of 32 patients; less than 15 mg/m2, six of six patients. Of the seven patients with headache during CRT, five developed the SS. The two patients (both of the less than 15 mg/m2 group) who did not develop the SS were the only cases treated with increased steroid doses at the onset of headache symptoms. Steroid coverage at a dose of greater than or equal to 15 mg/m2 during CRT appears to significantly reduce the incidence of acute radiation reactions and the SS. A prospective randomized study is planned to confirm these initial findings.

Preliminary results of alternating combination chemotherapy (CT) and hyperfractionated radiotherapy (HART) in advanced rhabdomyosarcoma (RMS)

In an attempt to reduce treatment (Rx) related acute toxicity and improve protocol compliance without compromising local control nor overall survival, a Phase II single arm pilot employing alternating intensive multiagent CT and radiotherapy (RT) in children with gross residual or metastatic RMS was begun at Memorial Sloan-Kettering Cancer Center (MSKCC) in 1984. Hyperfractionated radiotherapy (HART) was adopted to allow for timely delivery of both the RT and CT. From July, 1984 through July, 1986 12 patients (pts), aged 2 to 23 yr (median 10 yr) were enrolled on study. CT treatment was delivered over approximately 14 months and included 2 induction and 5 maintenance cycles. The induction CT consisted of two repetitive cycles of Vincristine, Dactinomycin, Cyclophosphamide, Adriamycin, Bleomycin, and Methotrexate; the maintenance CT included the same agents as reduced drug doses. The HART was delivered to the primary site during cycle I of induction at fractions of 150 cGy BID to a total dose of 5400 cGy in 2 courses of 3000 cGy and 2400 cGy, respectively. One hundred percent of patients completed the recommended dose of HART; 0% required unplanned interruptions of HART due to treatment toxicity. With a median follow-up (f/u) from diagnosis of surviving patients of 25 months (range 20-30), the local control rate is 83% (10/12 pts) and the overall survival, 58% (7/12 pts). The median time (MT) to any failure, local or metastatic, is 9 months (mo); and to death, 14 mo. Comparison of results with 12 historical controls with concomitant split-course standard fractionation RT (180-200 cGy/per fraction) and T6 CT during a MSKCC trial from 1975-1984, matched by site and stage of primary, revealed that 9 pts (75%) completed the recommended dose of RT, and 7 pts (58%) required interruptions of RT. With a median f/u of 78 mo (range 34-109), the local control rate was 75% (9/12 pts) and the overall survival, 42% (5/12 pts). The MT to any failure was 14 mo; and to death, 18 mo. These results indicate that the mode of alternating CT and HART (HART T6) as employed in this pilot study is well tolerated. It appears to offer a significant improvement in protocol compliance over previous protocols using concomitant CT and RT without any apparent compromise in local primary control or survival rates with a median f/u suggestive of adequate elapsed time for the appearance of most relapses and deaths in advanced RMS.

Is cranial radiation necessary for CNS prophylaxis in pediatric NHL?

The records of 95 consecutive children less than or equal to 21 years of age with previously untreated diffuse histology NHL registered in our protocols from 1978 to 1983 were reviewed. Seventy-nine patients were considered eligible for analysis. The histologic subtypes represented included lymphoblastic (LB) 37%; histiocytic (DHL) 29%; undifferentiated (DU) 19%; poorly differentiated (DPDL) 9%; and unclassified (UNHL) 6%. Distribution of the patients according to stage showed Stage I, 0%; Stage II, 11%; Stage III, 53%; Stage IV, 36%. Four different Memorial Hospital protocols for systemic chemotherapy were used (LSA2L2 73%; L10 9%; L17 10%; L17M 8%); however, the IT (intrathecal) chemotherapy was uniform (Methotrexate: 6.0-6.25 mg/M2 per treatment course) and was included in the induction, consolidation, and maintenance phases of all treatment protocols. Cranial radiation was included in the induction, consolidation, and maintenance phases of all treatment protocols. Cranial radiation was not included in the CNS prophylaxis program. The overall median time of follow-up was 43 months. The overall CNS relapse rate was 6.3%, however, the incidence of CNS lymphoma presenting as the first isolated site of relapse in patients in otherwise complete remission (minimum follow-up of 19 months with 97% of patients off treatment) was only 1/58 (1.7%). Our data suggests that IT chemotherapy when given in combination with modern aggressive systemic combination chemotherapy, and without cranial radiation appears to be a highly effective modality for CNS prophylaxis regardless of stage, histology, or bone marrow or mediastinal involvement. Therefore, with the commonly used aggressive combination chemotherapy for the management of all stage diffuse pediatric NHL, and the known increased risk of leukoencephalopathy with combination of cranial radiation and intensive systemic and intrathecal chemotherapy, we believe that cranial radiation may not be indicated for CNS prophylaxis in pediatric NHL.

Recurrent stage I endometrial carcinoma: results of treatment and prognostic factors

Recurrences of clinical Stage I endometrial carcinoma after initial treatment are rare. They are nonetheless a serious complication, uniformly associated with poor survival outcome. Between 1969-1980, 20 patients with clinical Stage I endometrial carcinoma were treated for recurrent tumor at the time of first relapse. Nonpapillary adenocarcinoma represented 70% of the primary tumors (pure adenocarcinoma, 50%; adenosquamous, 15%, clear cell, 5%) and papillary adenocarcinoma, 30%. The most common presenting symptom was vaginal bleeding, occurring in 95% of patients. The median time to recurrence after completion of primary treatment was 9.5 mo: Adenocarcinoma relapsed at a median time of 33 mo, adenosquamous, 6 mo and papillary adenocarcinoma, 4 mo. The vagina was the site of relapse in 65% of patients, the abdomen in 20%, the pelvis in 10% and the lung in 5%. Ninety-five percent of recurrences were treated with curative intent. Complications were seen in three patients, small bowel obstruction (2 pts) and vaginal vault necrosis (1 pt); however, these patients responded effectively to conservative treatment. Minimum follow-up of 4 years was available in 18 pts (90%). Actuarial 4 yr overall and NED survival was 50%, respectively, with a median survival of 39 mo to date. There have been no deaths from further recurrence of endometrial cancer beyond 39 mo. Significant prognostic factors for 4 year survival were 1) recurrence site--vagina, 82% (9/11 pts) vs extravagina, 0% (0/7 pts; median survival: 8 mo) [p = .0001]; and 2) histologic cell type--non-papillary carcinoma, 75% (9/12 pts) vs papillary adenocarcinoma, 0% (0/6 pts; median survival: 8 mo) [p = .002]. Our review suggests that: (1) Histology and site of relapse are important prognosticators of treatment outcome; (2) Long term survival may be achieved in vaginal recurrences with aggressive local treatment; and (3) There may be a role for multimodality ovarian type treatment in overall management of recurrent papillary adenocarcinoma, a cell type that appears to exhibit a tendency towards extrapelvic spread refractory to definitive loco-regional treatment.