Low-stage medulloblastoma: final analysis of trial comparing standard-dose with reduced-dose neuraxis irradiation.

PURPOSE: To evaluate prospectively the effects on survival, relapse-free survival, and patterns of relapse of reduced-dose (23.4 Gy in 13 fractions) compared with standard-dose (36 Gy in 20 fractions) neuraxis irradiation in patients 3 to 21 years of age with low-stage medulloblastoma, minimal postoperative residual disease, and no evidence of neuraxis disease. PATIENTS AND METHODS: The Pediatric Oncology Group and Children's Cancer Group randomized 126 patients to the study. All patients received posterior fossa irradiation to a total dose of 54 Gy in addition to the neuraxis treatment. Patients were staged postoperatively with contrast-enhanced cranial computed tomography, myelography, and CSF cytology. Of the registered patients, 38 were ineligible. RESULTS: The planned interim analysis that resulted in closure of the protocol showed that patients randomized to the reduced neuraxis treatment had increased frequency of relapse. In the final analysis, eligible patients receiving standard-dose neuraxis irradiation had 67% event-free survival (EFS) at 5 years (SE = 7.4%), whereas eligible patients receiving reduced-dose neuraxis irradiation had 52% event-free survival at 5 years (SE = 7.7%) (P =.080). At 8 years, the respective EFS proportions were also 67% (SE = 8.8%) and 52% (SE = 11%) (P =.141). These data confirm the original one-sided conclusions but suggest that differences are less marked with time. CONCLUSION: Reduced-dose neuraxis irradiation (23.4 Gy) is associated with increased risk of early relapse, early isolated neuraxis relapse, and lower 5-year EFS and overall survival than standard irradiation (36 Gy). The 5-year EFS for patients receiving standard-dose irradiation is suboptimal, and improved techniques and/or therapies are needed to improve ultimate outcome. Chemotherapy may contribute to this improvement.

Mucosa-associated lymphoid tissue lymphoma of the stomach: long term outcome after local treatment.

BACKGROUND: Although antibiotic therapy is emerging as effective initial treatment for patients with gastric lymphoma of mucosa-associated lymphoid tissue (MALT), there is a subset of patients for whom antibiotics are ineffective or inappropriate. Surgical resection can be curative, but total gastrectomy may be required for the eradication of all disease. To identify the optimal nonantibiotic therapy for early stage gastric MALT lymphoma, the authors retrospectively evaluated the Massachusetts General Hospital experience with gastric MALT lymphoma. METHODS: Disease patterns and treatment outcomes were retrospectively analyzed in data from 21 consecutive patients with gastric MALT lymphoma who were treated between 1978 and 1995 at the Massachusetts General Hospital. RESULTS: Sixteen patients were Stage IE, and 5 were in higher stages. Treatment consisted of resection with or without radiation or chemotherapy (14 patients), radiation alone (4 patients), or radiation plus chemotherapy (2 patients). Thirteen Stage IE patients received local therapy only. The 10-year actuarial relapse free survival rate for Stage IE patients was 93%, with 1 relapse among 15 treated patients. Because the patient who relapsed was treated successfully with chemotherapy, the 10-year cancer free survival was 100%. Overall survival for Stage IE patients was 93% at 5 years and 58% at 10 years, with no deaths from lymphoma. CONCLUSIONS: These data indicate that a high probability of long term remission can be achieved with only local treatment of patients with Stage I gastric MALT lymphoma. Preliminary results suggest that radiation therapy is well tolerated and effective and may well be the optimal nonantibiotic treatment for patients with localized gastric MALT lymphoma.

A blueprint for linking academic oncology and the community.

The processes by which academic medicine will train the next generation of physicians and develop new knowledge have brought to the forefront the relationship between academic medical centers and community hospitals and practitioners. Over the past thirteen years, the Harvard Medical School Joint Center for Radiation Therapy (JCRT) has developed an integrated radiation oncology program designed to serve simultaneously the needs of the community, teaching hospitals, medical schools, and faculty. The structure and function of this program are described here, as are the challenges posed by the highly competitive health care marketplace. We believe that long-term vision should guide short-term goals. The success of academic-community collaborative programs depends not only on the good will and vision of the participants but also on the medical administration, academic leadership, policy makers, and politicians who define the principles and rules by which cooperation within the health care industry occurs.

Stage IA-IIB Hodgkin's disease: management and outcome of extensive thoracic involvement.

PURPOSE: To examine the presentation, management, and outcome of patients with extensive intrathoracic involvement in early-stage Hodgkin's disease. PATIENTS AND METHODS: One hundred seventy-two patients with clinical Stage IA-IIB Hodgkin's disease and extensive intrathoracic involvement were studied. Extensive intrathoracic disease was defined as either large mediastinal adenopathy (LMA, defined as the width of the mass greater than one-third the maximum thoracic diameter, n = 154) or as extensive (> 10 cm) cephalocaudad intrathoracic disease that did not fulfill formal chest radiograph criteria for LMA (n = 18). Patients were divided into three groups based on staging and extent of treatment. Forty-seven patients were treated with radiation alone after a laparotomy (RT-lap), 47 patients received combined modality therapy after laparotomy (CMT-lap), and 78 patients were treated with combined modality therapy without staging laparotomy (CMT-no lap). MOPP was used in 82% of the CMT patients. Low-dose whole-cardiac RT was used in nearly 50% of patients treated either with RT or CMT. RESULTS: The 10-year actuarial freedom from relapse rates were 54% with RT alone and 88% with CMT (p = 0.001); overall survival rates were 84 and 89%, respectively (p = NS). The median time to relapse was only 17 months. Over 80% of relapses occurred within the first 3 years. The most common site of relapse in all patients was the mediastinum. Relapses below the diaphragm were rare, even in CMT patients who did not receive abdominal radiation treatment. The principal acute morbidity was symptomatic pneumonitis, which occurred in 29% of patients receiving any part of their chemotherapy after RT, compared to 13% if all the chemotherapy was given before RT and 11% if RT alone was administered. There was a low late risk of myocardial infarction (3%) in the two groups with the longest follow up (RT-lap, CMT-lap), but a higher risk of second malignancy in the CMT-lap group (21%) compared with the RT-lap group (2%). CONCLUSION: Extensive intrathoracic involvement is a distinctive presentation of early-stage HD that has a high relapse risk if treated with RT alone. The introduction of CMT has been associated with improvements in freedom from relapse. The low rate of peripheral relapse with CMT suggests that reductions in field size may be achievable. The use of low-dose whole-heart RT with modern techniques is not associated with a high risk of late cardiac complications and should be used in patients who present with extensive pericardial disease or cardiophrenic lymphadenopathy. The high rate of second malignancy in the CMT group with the longest follow-up suggests that careful long-term surveillance for such patients is warranted.

Results of a prospective randomized trial comparing standard dose neuraxis irradiation (3,600 cGy/20) with reduced neuraxis irradiation (2,340 cGy/13) in patients with low-stage medulloblastoma. A Combined Children's Cancer Group-Pediatric Oncology Group Study.

PURPOSE: To determine in a prospective randomized trial the effect on survival, progression-free survival, and patterns of relapse of a decrease in the neuraxis radiation dose from 3,600 cGy in 20 fractions to 2,340 cGy in 13 fractions in patients with newly diagnosed medulloblastoma between 3 and 21 years of age with low T stage (T1, T2 and T3A), minimal postoperative residual tumor, and no evidence of dissemination (M0). METHODS AND MATERIALS: Between June 1986 and November 1990, the Children's Cancer Group and the Pediatric Oncology Group randomized 126 patients in a two-arm study comparing the two different doses of neuraxis irradiation. In both arms, the posterior fossa received 5,400 cGy in 30 fractions. All patients were staged with myelography, postoperative lumbar cerebrospinal fluid cytology, and postoperative contrast-enhanced cranial computerized tomography to ensure no evidence of dissemination and no more than 1.5 cm3 residual tumor volume. Overall survival, progression-free survival, and patterns of recurrence were carefully monitored. Prospective endocrine and psychometric studies were performed to determine the benefit of decreasing the neuraxis radiation dose. RESULTS: Following an interim analysis at a median time on study of 16 months, the study was closed, since a statistically significant increase was observed in the number of all relapses as well as isolated neuraxis relapses in patients randomized to the lower dose of neuraxis radiation. CONCLUSIONS: In patients with newly diagnosed medulloblastoma considered to have a good prognosis on the basis of low T stage, minimal residual tumor after at least subtotal resection, and no evidence of dissemination after thorough evaluation, there is an increased risk of early relapse associated with lowering the dose of neuraxis radiation from 3,600 cGy in 20 fractions to 2,340 cGy in 13 fractions.

TGF alpha expression in meningioma--tumor progression and therapeutic response.

Little is known of the molecular genetic mechanisms contributing to meningioma tumor progression. We evaluated a total of 26 clinical cases of meningioma: twenty three patients with meningioma treated at our institution between 1978 and 1990 and three asymptomatic cases found initially at autopsy. In addition, histologically normal meninges obtained at post-mortem examination from 5 cases were evaluated. There were 13 men and 10 women in the patient group with a median age of 48.7 years, treated by surgery and/or irradiation. Median follow-up was 46 months (range 16-152 months). Archival cases and age-matched normal meningeal tissue obtained at autopsy during the same time period were obtained for study. Patients with TGF alpha scores greater than 3.0 were more likely to fail treatment and had lower overall survival times than those with immunostaining scores of 1 or 2. Three autopsy cases where meningioma had been silent clinically had overall staining scores of 0.75, while 10 samples of normal meninges harvested from 5 cases at autopsy had staining scores of 0. Two patients each underwent 3 surgeries for recurrent tumor, serial specimens showed increased TGF alpha expression over time, though all material from these procedures was consistent with the diagnosis of histologically benign meningioma.

Echo-planar MR cerebral blood volume mapping of gliomas. Clinical utility.

Neovascularization is a common phenomenon in gliomas. MR imaging cerebral blood volume (CBV) mapping utilizes ultrasfast echo-planar imaging and simultaneous use of gadolinium-based contrast material. To determine the utility of MR CBV mapping in the clinical evaluation of gliomas, we followed 15 patients with serial studies. This technique provided functional information that was not evident with conventional CT or MR imaging. Low-grade tumors demonstrated homogeneously low CBV, while high-grade tumors often showed areas of both high and low CBV. The maximum tumor CBV/white matter ratio was compared between low- (n = 3) and high-grade gliomas (n = 5) in patients without previous treatment and with histologic verification (n = 8) and was significantly higher in high-grade gliomas (p < 0.01). High CBV foci in nonenhancing tumor areas were present in 2 cases. The distinction between radiation necrosis and active tumor could be made correctly in 3 of 4 cases. The information provided by MR CBV mapping has the potential to be an adjunct in the clinical care of glioma patients.

Prognostic significance of proliferative indices in meningiomas.

The prognostic value of tumor proliferative indices in meningiomas was assessed by mitotic counts and by immunocytochemistry using a monoclonal antibody against the proliferating cell nuclear antigen (PCNA) (clone 19A2), a 36-kd nuclear protein involved in DNA synthesis. Sixty-three intracranial meningiomas were classified as benign (26), with atypical features (24) or as malignant (13). The patients included 24 men and 39 women, mean age 54.2 +/- 1.7 (mean +/- SEM) (range 15-78) at initial presentation. Twenty-four of the 63 primary tumors recurred locally, including 23.1% (6/26) of the benign, 37.5% (9/24) of the atypical, and 69.2% (9/13) of the malignant meningiomas. Among tumors that recurred, 1/9 (11%) of the atypical and 5/9 (55.5%) of the malignant tumors had had macroscopical total excision at the initial surgery. The mean interval to recurrence was 52 +/- 11.8 months. The mean progression-free follow-up period for patients without tumor recurrence was 82 +/- 8.5 months. Analysis of variance revealed that significant differences existed between tumor grades for both PCNA indices (1.16 +/- 0.29% for benign; 14.14 +/- 2.07% for atypical and 21.37 +/- 5.47% for malignant) and mitotic indices (total counts per ten high power fields) (0.08 +/- 0.05 for benign, 4.75 +/- 0.91 for atypical and 19.00 +/- 4.07 for malignant). Multivariate regression analysis indicated that mitotic index > 6 was the single most important factor (p < 0.05) for shorter disease-free interval. Age, sex and total surgical excision were not significant factors. PCNA index was a significant factor in the univariate model, but not in the multivariate model.(ABSTRACT TRUNCATED AT 250 WORDS)

Hodgkin disease, nodular sclerosis type. Implications of histologic subclassification.

BACKGROUND: The prognostic significance of the cellular composition of the nodules of Hodgkin disease, nodular sclerosis type (HDNS), is controversial. METHODS: Tumors from 79 patients with HDNS, who had a median follow-up time of 9.3 years, were studied. RESULTS: Based on British National Lymphoma Investigation criteria, 58 cases were classified as NSI (low-grade) and 21 as NSII (high-grade). The study included 24 male and 55 female patients, aged 10-57 years (mean, 27 years), who presented with Stage I (13 patients [12A, 1B]), Stage II (45 patients [40A, 5B]), or Stage III (21 patients [16A, 5B]) disease. Fifty-three patients had no relapse, 4 died of other causes, and 49 are in complete clinical remission. Twenty-six patients had progression of disease during therapy or relapsed and 17 were successfully salvaged. Overall length of survival was significantly shorter with NSII (P = 0.0001), extensive necrosis (P = 0.0034), high stage (P = 0.0058), and B symptoms (P = 0.030). Multivariate analysis showed that grade had the strongest effect on overall survival (P = 0.0042; hazard ratio = 10.19). The 5-year survival was 100% for NSI patients and 75% for NSII patients. Only B symptoms were significantly associated with risk of relapse after initial therapy (P = 0.030). For patients who relapsed, only histologic grade predicted subsequent disease-free survival (P = 0.0023; hazard ratio = 26.5). Five-year disease-free survival after first relapse was 94% for NSI patients and 11% for NSII patients. CONCLUSIONS: Patients with NSI disease who relapse have a more successful salvage and longer period of survival than do those with NSII disease. Histologic subclassification of HDNS appears clinically relevant, and consideration of histologic subtype may be important when planning therapy.

The role of radiotherapy in the treatment of subtotally resected benign meningiomas.

Thirty-six patients with benign meningioma were treated for primary or recurrent disease by subtotal resection and external beam irradiation from 1968-1986 at Massachusetts General Hospital. Comparison is made with 79 patients treated by subtotal surgery alone from 1962-1980. Progression-free survival for 17 patients irradiated after initial incomplete surgery was 88% at 8 years compared with 48% for similar patients treated by surgery alone (p = 0.057). 16 patients incompletely resected at time of first recurrence were irradiated and 78% were progression-free at 8 years while 11% of a similar group treated by surgery alone were progression free (p = 0.001). Long term overall survival was high and similar in both control and study groups. Two patients were irradiated at second recurrence and 1 patient at third recurrence. Twenty-five patients were treated with photons alone and have a median follow-up of 57 months, 6 patients have recurred at doses 45-60 Gy. Eleven patients were treated with combined 10 MV photons and 160 MV protons utilizing 3-D treatment planning. These patients have been followed for a median of 53 months and none have failed to date. Eight of 11 received 54-60.4 Gy and 3/11 greater than 64.48 Gy. Sex, age, pathology grade and score, surgery and timing of radiation therapy were not associated with significant differences in failure patterns within the irradiated study group (p less than 0.1). Complications have been seen in 6 irradiated patients.

Pineal tumors.

From 1972-1985, 30 patients with pineal area tumor have been treated with combinations of surgery and irradiation, no patient receiving chemotherapy as a primary recommendation. Patients ranged between 3 and 69 years of age, 18 were male and 12 female. In 18 patients a tissue diagnosis was made initially, 14 patients required shunts before definitive management and in 8 of the 14 tissue diagnosis was made at the same time. Eight patients had no surgical intervention at any time. Patients who have received irradiation have had whole brain irradiation, local field irradiation, or craniospinal irradiation. Two elderly patients died rapidly of their malignant processes before definitive treatment could be given. Only one patient with a non neoplastic lesion was seen. This was a bleed from an AVM with no underlying tumor. One patient with a pineocytoma was not irradiated. Overall, two-thirds of our patients are alive 5 to 15 years after treatment. There has been no surgical mortality and minimal morbidity from biopsy. Late effects of treatment include one patient with mild hearing loss and three patients with endocrinopathies amenable to medical treatment. We believe that tissue diagnosis allows optimal field design and dose recommendations to be made by the radiotherapist in addition to defining prognosis. In our experience, endodermal sinus tumor and pineoblastoma are highly malignant, and in view of their poor prognosis with conventional management consideration of more radical treatment with a possible role for chemotherapy is suggested.

A report of the consensus workshop panel on the treatment of brain metastases.

The consensus workshop on management of brain metastases focused on several broad issues concerning the patient with brain metastases. These issues included: What are the goals of treatment? Who should not be treated with radiation? What are the optimal treatment regimens? What are the areas of controversy where future investigations may be useful? The brain consensus panel attempted to develop guidelines for radiation oncologists through use of a decision tree format. Existing data to support management decisions, particularly those regarding radiation treatment, were reviewed. Data regarding cost of treatment as well as data regarding explicit quality of life measures were felt to be very important but were largely not available and therefore played little scientific role in the guideline development. Areas where controversy existed and where consensus could not be reached were identified.

Comparative treatment planning: proton vs. x-ray beams against glioblastoma multiforme.

The survival of patients with glioblastoma multiforme is extremely poor, the 5-year survival rate being almost zero. The cause of failure is almost exclusively local progression of tumor, the remainder is due to complications of treatment. Although this tumor is clearly radiation resistant, there is evidence of a dose response relationship. Using a thin slice CT scan of the entire head of a patient with glioblastoma multiforme, 3-dimensional radiation treatment plans were developed for treatment to a dose of 90 cobalt-Gray-equivalent (CGE). Dose distributions using protons were compared to those using x-rays. The results showed advantages for the proton beam technique. Namely the proton plan irradiated less non-target brain than the x-ray plan; this was especially so in the decrease of coverage of deep-seated structures. The volume of non-target brain that received more than 70 CGE was 175 ml for the x-ray plan and 94 ml for the proton plan. This study indicates that for a subpopulation of patients with glioblastoma multiforme, at least 90 CGE could be delivered with proton beam techniques to the target with only small volumes of normal brain structures receiving more than 70 CGE.

Radiation therapy in the management of early stage Hodgkin's disease presenting in later life.

This study examined the contention that although elderly patients with Hodgkin's disease have a worse prognosis overall than younger patients, a subgroup of older patients fit enough to be managed like younger patients can fare just as well. A retrospective analysis was made on 29 patients older than 60 years of age with Stage I and II Hodgkin's disease treated by radiation therapy alone. Fourteen of these patients were managed optimally, i.e., were adequately staged (defined by one or more of the following: laparotomy, computed tomography [CT] scan, and/or lymphangiogram), followed by radical radiation therapy (mantle or inverted-Y). The remaining 15 patients, because of their general medical condition, were managed suboptimally with limited staging and/or involved-field irradiation. None of the 14 patients managed optimally relapsed over a median of 4.75 years of follow-up compared with 10 of 15 patients in the suboptimal group. For the optimally managed versus suboptimally managed groups, the actuarial 5-year disease-free survival rates were 61% and 6%, respectively; the actuarial overall survival rates (death from all causes) were 61% and 19%, respectively; and the disease-specific survival rates were 100% and 39%, respectively. Only three of the patients irradiated radically had acute complications severe enough to warrant a break in treatment. In the opinion of the authors, those elderly patients able to tolerate adequate staging and radical radiation therapy can anticipate a high likelihood of cure.

Chemotherapeutic treatment of extensive optic pathway tumors in infants.

Two infants, ages 14 and 4 months, with extensive optic pathway tumors were treated with intensive chemotherapy called MADDOC: nitrogen mustard, doxorubicin, cis-platinum, dacarbazine, vincristine, and cyclophosphamide. The first child had hydrocephalus with an enhancing mass at the hypothalamus which followed the optic radiation to include the lateral geniculate body and medial temporal lobe. A v-p shunt was placed, and biopsy revealed a Grade II astrocytoma. One month later, the child developed malignant ascites. Intensive induction chemotherapy was then begun with cis-platinum 100 mg/m2 and cyclophosphamide 3 g/m2 for two initial cycles. The ascites resolved within one week, and chemotherapy was continued for 10 courses of the 6-drug MADDOC regimen. CT scans showed a gradual shrinkage of the tumor mass by approximately 70%. The enhancing areas continued to decrease in size through 20 months after completing MADDOC. The child has not received radiation and is well 4 years 7 months post diagnosis. The second infant had massive enlargement of the right optic nerve with an enhancing chiasmatic mass extending into the suprasellar space, hypothalamus, and brain stem. This infant was not biopsied; she also received induction MADDOC chemotherapy for 12 cycles. CT scans showed a definite decrease in the chiasmatic mass by the fifth cycle, with continued reduction by approximately 40% after 10 months. Twenty-three months from diagnosis there was asymptomatic evidence of tumor growth. The child is being treated with carboplatinum and remains ophthalmologically and radiographically stable 43 months from diagnosis.

Central nervous system neurocytoma and neuroblastoma in adults--report of eight cases.

The clinical features, pathologic findings and treatment courses of eight adults with central nervous system small-cell neuronal tumors were reviewed. Five patients had central neurocytomas, two patients central nervous system neuroblastomas, and one patient a neurocytoma-like spinal cord tumor. The neurocytomas were intraventricular, moderately cellular tumors with bland nuclei and perinuclear halos. Patients with neurocytoma were treated with surgery, radiation therapy, and/or chemotherapy, and have followed favorable clinical courses. The neuroblastomas were intraparenchymal, hypercellular tumors with necrosis and frequent mitoses. Patients with neuroblastomas were treated with surgery, radiation therapy and chemotherapy, with some clinical response, but overall poor survival. One of the two patients developed extracranial metastasis. The spinal cord tumor had histologic features of neurocytoma, and responded well to biopsy and radiation therapy. The cases are compared with the varieties of small-celled neuronal tumors described in the literature, and pathologic, histogenetic and treatment implications are discussed.

Conjunctival lymphoma: results and treatment with a single anterior electron field. A lens sparing approach.

Lymphoma of the conjunctiva is rare. It presents in older patients as a mass lesion and usually remains localized. Surgery is limited to biopsy, and radiation therapy is the definitive treatment of choice. The entire conjunctiva is treated. Relatively high doses (approximately 30 Gy) are required for local control, which may lead to cataract formation. Twelve patients with conjunctival lymphoma were treated at the Massachusetts General Hospital between 1979 and 1988. Ten of 12 patients presented with a unilateral lesion; 2 of 12 with bilateral lesions. Two of 12 patients were found to have systemic disease at the time of presentation. One patient developed conjunctival lymphoma 5 years after the diagnosis of generalized disease. Using electron beam, all patients were treated with a single anterior circular field to total doses ranging from 24 Gy to 30 Gy delivered in 8 to 16 fractions over 9 to 20 days. In all cases, the lens was shielded by a specially designed plastic contact lens bearing a 12 mm diameter lead shield. The lens dose was determined at varying depths beneath the shield for 6 MeV and 9 MeV electron beams and ranged from a minimum of 5% to an absolute maximum of 18% of the total dose delivered to the tumor. Local control was maintained in all patients with follow-up to 9 1/2 years. One patient relapsed distantly 3 years after treatment. One of 12 patients died of systemic disease 4 years after treatment of the ocular lesion. Two patients developed cataracts 4 and 5 years after treatment; one had bilateral cataract, although only one eye had been treated. Both patients were over 75 years old. In both cases, the cataracts were felt to be senile cataracts which are ophthalmologically and radiographically distinguishable from radiation induced lesions.

Fractionated proton radiation therapy of cranial and intracranial tumors.

Since 1973 fractionated proton radiation therapy has been used in the treatment of malignant disease. Protons have favorable physical characteristics that yield dose distributions superior to those of photons in certain clinical situations. As of December 31, 1987, 1,678 patients had been treated. Of these, 110 had chordomas or low-grade chondrosarcomas of the base of skull. The first 68 patients have a minimum follow-up of 17 months. The median dose was 69 Cobalt Gray Equivalent (CGE). (CGE is the dose in proton Gray multiplied by 1.1, which is the relative biological effectiveness for protons relative to 60Cobalt.) The actuarial 5-year local control rate is 82%, and the disease-free survival rate is 76%. Thirteen patients with meningiomas have been treated, following subtotal resection. The median dose was 59.4 CGE. With a median follow-up of 26 months, no patient has had tumor progression. In addition, nine patients with gliomas and 12 with craniopharyngiomas have been treated.

Immunohistologic features predict clinical behavior of orbital and conjunctival lymphoid infiltrates.

The natural history of lymphoid infiltrates of the orbit and conjunctiva is poorly understood. To determine if immunohistologic features could predict clinical outcome, these features were evaluated in 61 patients with orbital and conjunctival lymphoid infiltrates, including 44 patients with lesions localized to one or both orbits or conjunctivae. Using histologic criteria, 20 infiltrates were classified as malignant lymphoma (cytologically atypical), 14 cases were classified as benign (follicular hyperplasia or inflammatory pseudotumor), and 27 infiltrates (44%) were dense infiltrates of small lymphocytes without cytologic atypia, which could not be confidently classified as benign or malignant (histologically indeterminate). Based on expression of monotypic immunoglobulin, 20 of these cases were reclassified as small lymphocytic lymphoma. For all cases, monotypic immunoglobulin expression correlated with reduced survival (P less than .05) and increased likelihood of dissemination (P less than .001). Monotypic immunoglobulin expression also correlated with increased risk of dissemination for all histologically indeterminate (small lymphocytic) infiltrates (P less than .05). Separate analysis of the localized infiltrates showed that monotypic immunoglobulin expression significantly correlated with an increased likelihood of dissemination for all 44 cases (P less than .005) and for the 22 histologically indeterminate lesions (P = .06). For the localized small lymphocytic infiltrates, monotypic immunoglobulin expression conferred a 50% risk of dissemination. In contrast, no patients with polytypic small lymphocytic infiltrates have disseminated, although one lesion locally recurred 30 months later as lymphoma. Thus, monotypic immunoglobulin expression significantly correlates with reduced patient survival and increased risk of dissemination by orbital and conjunctival lymphoid infiltrates. Monotypic small lymphocytic infiltrates without cytologic atypia behave as do low grade B-cell malignant lymphomas with a significant risk of dissemination and an indolent clinical course. The relationship of polytypic small lymphocytic infiltrates to lymphoma remains to be determined.