F. Ann Hayes

Age‐linked prognostic categorization based on a new histologic grading system of neuroblastomas. A clinicopathologic study of 211 cases from the pediatric oncology group

Histologic sections (minimum of four sections per patient) from 211 patients with neuroblastoma were reviewed. The tumors were resected before therapy, which was standardized according to age and stage. Low mitotic rate (MR) (≤ ten per ten high‐power fields) and calcification emerged as the most significant prognostic features after statistical analysis by stepwise log‐rank tests (P < 0. 0001 and P = 0. 0065, respectively). Histologic Grades 1, 2, and 3 were defined on the basis of the presence of both, any one, or none of these two prognostic features, respectively (Grade 3 had absence of low MR, i.e., these tumors had high MR [> ten per ten high‐power fields]). Statistically significant differences in survival were observed in the grades after adjusting for age and stage (P < 0. 001). The degree of differentiation, although significant by itself, was no longer significant after adjusting for the grades, Age groups (≤ 1 versus > 1 year of age), which also emerged as an independent prognostic feature (P < 0. 001), were linked with the grades to define two risk groups as follows: (1) a low‐risk (LR) group consisting of patients in both age groups with Grade 1 tumors and patients 1 year of age or younger with Grade 2 tumors and (2) a high‐risk (HR) group consisting of patients older than 1 year of age with Grade 2 tumors and patients in both age groups with Grade 3 tumors. The difference in survival between LR (160 cases) and HR groups (51 cases) was statistically significant (P < 0. 001). Concordance between these LR and HR groups and the Shimada classification was observed in 84% of cases. The new histologic grading system has the following advantages: (1) use of familiar terminology and histologic features in the grading system and (2) relative ease of assessment because the degree of differentiation does not need to be determined. The grading system should be tested on a new data set with an appropriate histologic sample of similar size to confirm these results.

EWING'S SARCOMA: LOCAL TUMOR CONTROL AND PATTERNS OF FAILURE FOLLOWING LIMITED-VOLUME RADIATION THERAPY

Sixty children with localized osseous Ewings sarcoma were treated between 1978 and 1988 with induction chemotherapy (cyclophosphamide, adriamycin), irradiation and/or surgery, and 10 months of maintenance chemotherapy (cyclophosphamide, adriamycin, dactinomycin, vincristine). Following induction chemotherapy, 43 patients received primary radiation therapy to limited radiation volumes defined by post-chemotherapy residual soft tissue tumor extension and initial osseous tumor extent. Irradiation was defined as low dose at 30-36 Gy (median 35 Gy) for 31 cases with objective response to induction chemotherapy and high dose at 50-60 Gy (median 50.4 Gy) for 12 patients with poor response to induction chemotherapy or with tumors greater than or equal to 8 cm. Overall event-free survival at 5 years is 59% and local tumor control is 68%. Initial failures have been local (12), simultaneous local and distant failures (7), and distant (6). In the surgical resection group, 14 patients had complete resection without radiation therapy, and 3 patients had microscopic residual plus 35-41 Gy; 100% local control has been maintained. In 43 patients with primary radiation therapy group, local tumor control is 58% (p = .004). Despite limited radiation volume, 18/19 local failures occurred centrally within the bone, well within the radiation volume. Imaging response to induction chemotherapy predicted local tumor control in the radiation therapy group: 62% with complete response/partial response versus 17% with no response/progressive disease (p less than 0.01). Local tumor control related strongly to primary tumor size in the radiation therapy group; among 31 cases receiving 35 Gy, local tumor control is 90% for lesions less than 8 cm versus 52% for tumors greater than or equal to 8 cm (p = .054). The central pattern of local failure in this experience suggests the effectiveness of limited radiation volume. The overall local tumor control rate following the tested dose level of 35 Gy appears to be inadequate, although results in selected cases with tumors less than 8 cm in greatest tumor dimension indicate potential efficacy in a yet limited experience.

Sequential cyclophosphamide and doxorubicin for induction of complete remission in children with disseminated neuroblastoma

When 70 children with disseminated neuroblastoma were treated for 4 months with a course of cyclophosphamide and doxorubicin (Adriamycin) 35 (52%) of 68 assessable patients achieved complete remissions, 13 had partial responses, and 20 had either less than a 50% regression of tumor or evidence of tumor progression. Of the 35 children with complete responses, three received local irradiation, but for the remainder, chemotherapy alone provided rapid control of tumor growth. Drug‐induced toxicity caused the hospitalization of four of 17 infants and five of 52 older children, but was related to only two of the eight deaths that occurred during treatment. Infants less than 1 year of age were more likely to achieve a complete remission than were older children. That 52% of these patients achieved complete remissions represents a marked improvement over the 22% response rate for patients seen between 1962 and 1974. Of 35 patients with complete responses, 15 have survived, contrasted with only four of 33 patients achieving less than a complete response. Median survival for patients achieving a partial response was 14.4 months, compared to 8.4 months for patients not achieving a response. A kinetically based sequence of chemotherapy seems to produce a high rate of complete remissions in children with disseminated neuroblastoma. To consolidate this gain into curative treatment will require a continuation therapy to prevent the emergence of drug‐resistant disease.

Clinical evaluation of sequentially scheduled cisplatin and vm26 in neuroblastoma: Response and toxicity

Cis‐dichlorodiammineplatinum (CDDP) and VM26, both of which have been proven efficient in treating neuroblastoma, were combined in a sequential schedule and administered to 22 children with disseminated neuroblastomas resistant to treatment with cyclophosphamide and doxorubicin (Adriamycin). During this same study, 14 children were prospectively evaluated for the effect of CDDP on magnesium metabolism and the effect of the induced hypomagnesemia on parathyroid function. Complete or partial tumor responses were achieved in six and nine cases, respectively and were of prolonged duration (longer than six months) in eight of the 15 responding. It was also shown that CDDP‐induced hypomagnesemia is the result of excessive renal loss and is severe enough to interfere with the normal parathormone response to hypocalcemia.

Recommendations for modification of terminology of neuroblastic tumors and prognostic significance of Shimada classification. A clinicopathologic study of 213 cases from the Pediatric Oncology Group.

To develop consistency in terminology and pathologic criteria, the authors reviewed the literature and 213 cases of neuroblastic tumors (NT) registered with Pediatric Oncology Group (POG) protocols 8104 and 8441. The patients were given standardized therapy stratified according to POG stage and patient age, and four or more histologic sections of primary tumor resected before therapy were available in each of these 213 cases. All stages were represented. The recommended nomenclature combines conventional terms and criteria with those used by Bove and McAdams and Shimada et al. The main features of the recommended nomenclature are as follows: (1) the terms neuroblastoma (NB) and ganglioneuroblastoma (GNB) are retained instead of stroma‐poor NB and stroma‐rich NB, recommended by Shimada et al.; (2) undifferentiated NB is considered a subtype separate from poorly differentiated NB; and (3) the term GNB is used only when there is a predominant ganglioneuromatous component admixed with the minor neuroblastomatous component. With the use of these criteria and terms, the Shimada classification was determined in the 213 cases. The results showed that, even after stratification for age, POG stage, and primary site, there is a statistically significant difference in survival rate between favorable histologic and unfavorable histologic prognostic subgroups. The authors recommend that definitive prognostic categorization of an NT according to Shimada classification should be done only when adequate histologic material is available from a primary tumor resected before any other therapy. Categorization done on histologic material from small biopsy specimens, previously treated primary tumors, or meta‐static sites should be considered tentative.

Surgicopathologic staging of neuroblastoma." Prognostic significance of regional lymph node metastases

Of 254 children with neuroblastoma treated at St. Jude Childrens Research Hospital, 102 (40%) had clinically localized tumors. Using a surgicopathologic staging system, 66 of these 102 children had localized tumor and 36 had tumor dissemination to regional lymph nodes. Survival of these two groups has been markedly different; 57 of 66 (87%) with localized tumor survive, compared with 11 of 35 (33%) with node dissemination, five of whom are less than two years from diagnosis. Our data indicate that the prognosis for children with neuroblastoma metastatic only to regional lymph nodes is no different from that of patients of similar age with widely disseminated tumor.

Thoracic neuroblastoma: A pediatric oncology group study

Ninety-six patients with thoracic neuroblastoma were studied in a prospective fashion. Median age at presentation was 0.9 years. Forty-eight percent of the patients presented with stage A disease, 20% stage B, 13% stage C, 17% stage D, and 2% stage DS. Seventy-five patients have been followed for greater than 4 years. A posterior mediastinal mass was diagnosed on incidental chest roentgenograms performed for nontumor-related symptoms in 49% of the cases. Sixteen percent of the cases presented with neurological symptoms and 14% of the patients presented with acute respiratory distress. Urinary catecholamines were elevated in 76% of the cases. Complete surgical resection was carried out in 47% of the cases, while incomplete resection or biopsy was performed in 45%. No operation was performed in 3 patients. Minor surgical complications occurred in 20% of the patients, and 3% of the patients had significant perioperative complications. One patient died as a complication of therapy. Overall actuarial survival was 88% at 4 years. This study confirms the favorable outcome in children with mediastinal neuroblastoma. The basic biology of thoracic neuroblastomas seems to differ from that of other sites in that the majority of patients present at a younger age with localized disease or regional lymph node metastases, and have an improved survival even after correcting for age and stage. While complete excision is recommended, if possible, radical surgical procedures are not indicated since an excellent prognosis is associated with combined modality therapy.

Immunohistochemical expression of neuron-specific enolase and leu 7 in Ewing's sarcoma of bone

Neural features have been documented in a series of small round cell tumors of bone, previously classified as Ewings sarcoma of bone (ESB), using light microscopic, ultrastructural, immunohistochemical, and in vitro techniques. To date, correlation of the presence of these features in ESB with clinical outcome has not been performed. The authors investigated the clinical relevance of positivity to antibodies against neuron‐specific enolase (NSE) and Leu 7 (HNK 1) using the avidin‐biotin complex technique in 40 cases of ESB seen at St. Jude (Memphis, TN) during the period 1968 to 1986. Twenty‐three cases (58%) were positive for NSE and/or Leu 7. The median disease‐free survival of patients with localized tumors and NSE and/or Leu 7 positivity was 3.8 years compared to a median disease‐free survival of 1.6 years for those without these markers and localized disease. Using the log‐rank test, a statistically significant difference was shown in these two groups of patients (P = 0.049). The authors conclude that in this relatively small series ESB commonly expresses NSE and/or Leu 7 and that the presence of these markers is of no prognostic significance overall, but that in patients with localized ESB the presence of NSE and/or Leu 7 by immunostaining may be a favorable prognostic indicator.

Brain tumors in the very young child: Postoperative chemotherapy in combined-modality treatment

Twelve consecutively diagnosed children with brain tumors, ages 7 to 27 months, were treated by a combined‐modality approach featuring aggressive surgical resection followed by chemotherapy and delayed irradiation. Patients received multiple clinical neurologic examinations and psychological evaluations, as well as diagnostic imaging studies, to monitor the efficacy of chemotherapy and toxic effects of therapy. Six of the eight children with residual tumor evident postoperatively on computed tomography scans had objective responses to chemotherapy. The efficacy of chemotherapy was further demonstrated by the lack of disease progression for 7 months or longer in eight children, seven of whom remain free of tumor 19 to 57 months (median, 28 months) from the date of diagnosis. The neurologic status of ten patients improved during treatment. Developmental progress was normal in two, low average in three, and frankly deficient in four of ten children formally evaluated. These results indicate that postoperative chemotherapy, added to a regimen of surgical resection and delayed irradiation, prolong survival with only minimal short term neurotoxicity in the majority of very young children with malignant brain tumors.

Primary extracranial neuroblastoma with central nervous system metastases characterization by clinicopathologic findings and neuroimaging

The authors report the clinicopathologic and neuroimaging findings in ten children with primary abdominal or thoracic neuroblastoma who relapsed in the central nervous system (CNS) without evidence of concurrent intracranial extension from adjacent bone, dura, or dural sinus metastases. At diagnosis, the patients ranged in age from 0.3 to 4.5 years (median, 2 years). Their times to CNS relapse ranged from 2 to 34 months from diagnosis. In seven patients the relapse occurred from 1 to 14 months after elective discontinuation of therapy. In four patients, the CNS relapse was the primary (isolated) adverse event. Four patients could not be treated at the time of relapse, and they died within 7 days of progressive CNS disease. In the remaining group, craniospinal irradiation with or without administration of a platinum compound and an epipodophyllotoxin caused complete CNS remissions lasting 4, 5, 16, and 62+ months. Neuroimaging and autopsy findings indicated that cerebrospinal fluid is the major pathway for neuraxis dissemination by neuroblastoma cells. There was no evidence of dural penetration in any patient. The possibility of relapse in the neuraxis should be considered for any patient with neuroblastoma who has neurologic deterioration. A combination of craniospinal radiation and administration of a platinum compound and an epipodophyllotoxin will induce complete responses in some patients with neuraxis involvement by neuroblastoma, but the risk of subsequent failure outside the CNS remains high.