Faculty Bibliography

Purpose/UNASSIGNED:COVID-19 profoundly affected the United States, with New York City rapidly becoming the epicenter of the disease. Patients with cancer represent a vulnerable population in this pandemic, with data suggesting a higher risk for severe events and unfavorable outcomes. Timely identification of COVID-19 in patients with cancer has been thwarted by the limited availability of outpatient testing for SARS-CoV-2. Chest computed tomography (CT) plays a major role in the identification of COVID-19 pneumonia, with radiologic hallmarks including bilateral, peripheral ground-glass opacities (GGOs) and consolidation. Patients with cancer undergoing radiation therapy (RT) commonly have daily cone beam computed tomography (CBCT) obtained for image-guided RT, and such imaging frequently includes the chest. Methods and Materials/UNASSIGNED:We retrospectively reviewed the CBCT scans of an initially asymptomatic patient undergoing image-guided RT for breast cancer who developed COVID-19 symptoms during the second week of RT. Lung windows of daily CBCT scans were reviewed with diagnostic radiology to survey for changes consistent with COVID-19. Diagnostic CT scans at the time of recovery were obtained and compared with the CBCTs. Results/UNASSIGNED:Five consecutive CBCT scans were retrospectively reviewed. Bilateral, peripheral GGOs were noted on the fourth and fifth CBCT scans in the 2 days before symptom onset. CBCT on the day of RT resumption demonstrated substantial worsening of the GGO compared with scans obtained during the asymptomatic phase. Diagnostic CTs demonstrated bilateral, peripheral GGOs and mediastinal lymphadenopathy, findings suggesting COVID-19 pneumonitis. Repeat diagnostic CT 3 days later showed improved pulmonary findings, and the patient resumed RT without incident. Conclusions/UNASSIGNED:Familiarity with typical CT changes of COVID-19 pneumonitis may allow for early detection in cancer patients undergoing CBCT for RT treatment. Prompt review of the lung windows is recommended to identify such changes, with the hope that presymptomatic diagnosis leads to expedited patient management, improved outcomes, and a reduction of inadvertent COVID-19 dissemination.

OBJECTIVE The incidence of brain metastases is increasing with improved systemic therapies, many of which have a limited impact on intracranial disease. Stereotactic radiosurgery (SRS) is a first-line management option for brain metastases. The purpose of this study was to determine if there is a threshold tumor size below which local control (LC) rates approach 100%, and to relate these findings to the use of routine surveillance brain imaging. METHODS From a prospective registry, 200 patients with 1237 brain metastases were identified who underwent SRS between December 2012 and May 2015. The median imaging follow-up duration was 7.9 months, and the median margin dose was 18 Gy. The maximal diameter and volume of tumors were measured. Histological analysis included 96 patients with non-small cell lung cancers (NSCLCs), 40 with melanoma, 35 with breast cancer, and 29 with other histologies. RESULTS Almost 50% of brain metastases were NSCLCs and commonly measured less than 6 mm in maximal diameter or 70 mm3 in volume. Thirty-three of 1237 tumors had local progression at a median of 8.8 months. The 1- and 2-year actuarial LC rates were 97% and 93%, respectively. LC of 100% was achieved for all intracranial metastases less than 100 mm3 in volume or 6 mm in diameter. Patients whose tumors at first SRS were less than 10 mm maximal diameter or a volume of 250 mm3 had improved overall survival. CONCLUSIONS SRS can achieve LC rates approaching 100% for subcentimeter metastases. The earlier initial detection and prompt treatment of small intracranial metastases may prevent the development of neurological symptoms and the need for resection, and improve overall survival. To identify tumors when they are small, routine surveillance brain imaging should be considered as part of the standard of care for lung, breast, and melanoma metastases. CLASSIFICATION OF EVIDENCE Type of question: prognostic; study design: retrospective cohort; evidence: Class II.

Leptomeningeal disease (LMD) is well described in patients with brain metastases, presenting symptomatically in approximately 5% of patients. Conventionally, the presence of LMD is an indication for whole brain radiation therapy (WBRT) and not suitable for stereotactic radiosurgery (SRS). The purpose of the study was to evaluate the local control and overall survival of patients who underwent SRS to focal LMD. We reviewed our prospective registry and identified 32 brain metastases patients with LMD, from a total of 465 patients who underwent SRS between 2013 and 2015. Focal LMD was targeted with SRS in 16 patients. The median imaging follow-up time was 7 months. The median volume of LMD was 372 mm3 and the median margin dose was 16 Gy. Five patients underwent prior WBRT. Histology included non-small cell lung (8), breast (5), melanoma (1), gastrointestinal (1) and ovarian cancer (1). Follow-up MR imaging was available for 14 patients. LMD was stable in 5 and partially regressed in 8 patients at follow-up. One patient had progression of LMD with hemorrhage 5 months after SRS. Seven patients developed distant LMD at a median time of 7 months. The median actuarial overall survival from SRS for LMD was 10.0 months. The 6-month and 1-year actuarial overall survival was 60% and 26% respectively. Six patients underwent WBRT after SRS for focal LMD at a median time of 6 months. Overall, focal LMD may be may be treated successfully with radiosurgery, potentially delaying WBRT in some patients.

We prospectively addressed whether EGFR and KRAS mutations, EML4-ALK, ROS1 and RET rearrangements, or wild-type (WT), affects radiosurgery outcomes and overall survival (OS) in non-small cell lung cancer (NSCLC) patients with brain metastases (BM). Of 326 patients with BM treated in 2012-2014 with Gamma Knife radiosurgery (GKRS), 112 NSCLC patients received GKRS as their initial intracranial treatment. OS, intracranial progression-free survival, and time to intracranial failure were determined. Univariate and multivariate analysis were performed to determine factors affecting OS. Toxicity of treatment was evaluated. Median follow-up was 9 months. Patients with EGFR mutant BM had improved survival compared to WT. Median time to development of BM was higher in EGFR mutant patients, but this difference was not significant (2.2 vs 0.9 months; p = 0.2). Median time to distant brain failure was independent of EGFR mutation status. Karnofsky performance status (KPS), non-squamous histopathology, targeted therapy, systemic disease control, EGFR mutation, and low tumor volume were predictive of increased OS on univariate analysis. KPS (p = 0.001) and non-squamous histopathology (p = 0.03) continued to be significant on multivariate analysis. Patients with EGFR mutant BM underwent salvage treatment more often than those without (p = 0.04). Treatment-related toxicity was no different in patients treated with GKRS combined with targeted therapies versus GKRS alone (5 vs 7 %, p = 0.7). Patients with EGFR mutant BM had improved survival compared to a WT cohort. Intracranial disease control following radiosurgery was similar for all tumor subtypes. Radiosurgery is effective for BM and concurrent treatment with targeted therapy appears to be safe.

BACKGROUND: Previous reports have proposed an association between traumatic brain injury (TBI) and subsequent glioblastoma (GBM) formation. METHODS: We used literature searches and radiographic evidence from two patients to assess the possibility of a link between TBI and GBM. RESULTS: Epidemiological studies are equivocal on a possible link between brain trauma and increased risk of malignant glioma formation. We present two case reports of patients with GBM arising at the site of prior brain injury. CONCLUSION: The hypothesis that TBI may predispose to gliomagenesis is disputed by several large-scale epidemiological studies, but supported by some. Radiographic evidence from two cases presented here suggest that GBM formed at the site of brain injury. We propose a putative pathogenesis model that connects post-traumatic inflammation, stem and progenitor cell transformation, and gliomagenesis.