Collagen deposition within brain metastases is associated with leptomeningeal failure after cavity-directed radiosurgery.

Background: Leptomeningeal failure (LMF) represents a devastating progression of disease following resection of brain metastases (BrM). We sought to identify a biomarker at time of BrM resection that predicts for LMF using mass spectrometry-based proteomic analysis of resected BrM and to translate this finding with histochemical assays. Methods: We retrospectively reviewed 39 patients with proteomic data available from resected BrM. We performed an unsupervised analysis with false discovery rate adjustment (FDR) to compare proteomic signature of BrM from patients that developed LMF versus those that did not. Based on proteomic analysis, we applied trichrome stain to a total of 55 patients who specifically underwent resection and adjuvant radiosurgery. We used competing risks regression to assess predictors of LMF. Results: Of 39 patients with proteomic data, FDR revealed type I collagen-alpha-1 (COL1A1, P = .045) was associated with LMF. The degree of trichrome stain in each block correlated with COL1A1 expression (ß = 1.849, P = .001). In a cohort of 55 patients, a higher degree of trichrome staining was associated with an increased hazard of LMF in resected BrM (Hazard Ratio 1.58, 95% CI 1.11-2.26, P = .01). Conclusion: The degree of trichrome staining correlated with COL1A1 and portended a higher risk of LMF in patients with resected brain metastases treated with adjuvant radiosurgery. Collagen deposition and degree of fibrosis may be able to serve as a biomarker for LMF.

A transformer-based deep-learning approach for classifying brain metastases into primary organ sites using clinical whole-brain MRI images.

Treatment decisions for brain metastatic disease rely on knowledge of the primary organ site and are currently made with biopsy and histology. Here, we develop a deep-learning approach for accurate non-invasive digital histology with whole-brain magnetic resonance imaging (MRI) data. Contrast-enhanced T1-weighted and fast spoiled gradient echo brain MRI exams (n = 1,582) were preprocessed and input to the proposed deep-learning workflow for tumor segmentation, modality transfer, and primary site classification into one of five classes. Tenfold cross-validation generated an overall area under the receiver operating characteristic curve (AUC) of 0.878 (95% confidence interval [CI]: 0.873,0.883). These data establish that whole-brain imaging features are discriminative enough to allow accurate diagnosis of the primary organ site of malignancy. Our end-to-end deep radiomic approach has great potential for classifying metastatic tumor types from whole-brain MRI images. Further refinement may offer an invaluable clinical tool to expedite primary cancer site identification for precision treatment and improved outcomes.

Multiparametric radiomic tissue signature and machine learning for distinguishing radiation necrosis from tumor progression after stereotactic radiosurgery.

BACKGROUND: Stereotactic radiosurgery (SRS) may cause radiation necrosis (RN) that is difficult to distinguish from tumor progression (TP) by conventional MRI. We hypothesize that MRI-based multiparametric radiomics (mpRad) and machine learning (ML) can differentiate TP from RN in a multi-institutional cohort. METHODS: Patients with growing brain metastases after SRS at 2 institutions underwent surgery, and RN or TP were confirmed by histopathology. A radiomic tissue signature (RTS) was selected from mpRad, as well as single T1 post-contrast (T1c) and T2 fluid-attenuated inversion recovery (T2-FLAIR) radiomic features. Feature selection and supervised ML were performed in a randomly selected training cohort (N = 95) and validated in the remaining cases (N = 40) using surgical pathology as the gold standard. RESULTS: One hundred and thirty-five discrete lesions (37 RN, 98 TP) from 109 patients were included. Radiographic diagnoses by an experienced neuroradiologist were concordant with histopathology in 67% of cases (sensitivity 69%, specificity 59% for TP). Radiomic analysis indicated institutional origin as a significant confounding factor for diagnosis. A random forest model incorporating 1 mpRad, 4 T1c, and 4 T2-FLAIR features had an AUC of 0.77 (95% confidence interval [CI]: 0.66-0.88), sensitivity of 67% and specificity of 86% in the training cohort, and AUC of 0.71 (95% CI: 0.51-0.91), sensitivity of 52% and specificity of 90% in the validation cohort. CONCLUSIONS: MRI-based mpRad and ML can distinguish TP from RN with high specificity, which may facilitate the triage of patients with growing brain metastases after SRS for repeat radiation versus surgical intervention.

Cisplatin/5-Fluorouracil (5-FU) Versus Carboplatin/Paclitaxel Chemoradiotherapy as Definitive or Pre-Operative Treatment of Esophageal Cancer.

Purpose To determine the efficacy and toxicity of two standard chemotherapy regimens used concurrent with radiation for the treatment of esophageal cancer: cisplatin/5-fluorouracil (5-FU) and carboplatin/paclitaxel. Materials and methods We prospectively reviewed records of 364 patients with histologically confirmed stage I to IVA esophageal cancer receiving chemoradiotherapy (CRT) with or without resection. All patients received surgical evaluation and imaging at presentation as well as following completion of their course of CRT. Treatment and prognostic variables were compared across the two chemotherapy regimens. Results We identified 261 patients treated with concurrent carboplatin/paclitaxel (n = 133) or cisplatin/5-FU (n = 128). Weight loss during CRT was lower in patients receiving carboplatin/paclitaxel (median: 7.0 pounds; 4.1% body weight) vs. cisplatin/5-FU (median: 11.0 pounds; 6.5% body weight) (p < 0.01). In 117 patients receiving trimodality therapy, post-operative death rates within one month of resection were similar. Pathologic complete response was better with carboplatin/paclitaxel vs. cisplatin/5-FU, 29.6% vs. 21.8% (p = 0.03), respectively. In the multivariable analysis, there was no association between chemotherapy regimen and overall survival (OS) or progression-free survival (PFS), though there was a trend toward improved OS with carboplatin/paclitaxel with a HR = 0.75 (p = 0.08). Further analysis revealed that trimodality therapy and stage were predictors for improved OS and PFS while female gender and grade predicted for improved PFS. Conclusions Carboplatin/paclitaxel was associated with decreased weight loss and improved pathologic response for trimodality patients when compared to cisplatin/5-FU. We observed no differences in OS, PFS, or post-operative death by chemotherapy regimen for both the entire cohort and trimodality patients.

Community economic factors influence outcomes for patients with primary malignant glioma.

BACKGROUND: Community economics and other social health determinants influence outcomes in oncologic patient populations. We sought to explore their impact on presentation, treatment, and survival in glioma patients. METHODS: A retrospective cohort of patients with glioma (World Health Organization grades III-IV) diagnosed between 1999 and 2017 was assembled with data abstracted from medical record review. Patient factors included race, primary care provider (PCP) identified, marital status, insurance status, and employment status. Median household income based on zip code was used to classify patients as residing in high-income communities (HICs; ie, above the median state income) or low-income communities (LICs; ie, below the median state income). The Kaplan-Meier method was used to assess overall survival (OS); Cox proportional hazards regression was used to explore associations with OS. RESULTS: Included were 312 patients, 73% from LICs. Survivors residing in LICs and HICs did not differ by age, sex, race, tumor grade, having a PCP, employment status, insurance, time to presentation, or baseline performance status. Median OS was 4.1 months shorter for LIC patients (19.7 vs 15.6 mo; hazard ratio [HR], 0.75; 95% CI: 0.56-0.98, P = 0.04); this difference persisted with 1-year survival of 66% for HICs versus 61% for LICs at 1 year, 34% versus 24% at 3 years, and 29% versus 17% at 5 years. Multivariable analysis controlling for age, grade, and chemotherapy treatment showed a 25% lower risk of death for HIC patients (HR, 0.75; 95% CI: 0.57-0.99, P < 0.05). CONCLUSIONS: The economic status of a glioma patient's community may influence survival. Future efforts should investigate potential mechanisms such as health care access, stress, treatment adherence, and social support.

Multi-institutional validation of brain metastasis velocity, a recently defined predictor of outcomes following stereotactic radiosurgery.

INTRODUCTION: Brain metastasis velocity (BMV) is a prognostic metric that describes the recurrence rate of new brain metastases after initial treatment with radiosurgery (SRS). We have previously risk stratified patients into high, intermediate, and low-risk BMV groups, which correlates with overall survival (OS). We sought to externally validate BMV in a multi-institutional setting. METHODS: Patients from nine academic centers were treated with upfront SRS; the validation cohort consisted of data from eight institutions not previously used to define BMV. Patients were classified by BMV into low (<4 BMV), intermediate (4-13 BMV), and high-risk groups (>13 BMV). Time-to-event outcomes were estimated using the Kaplan-Meier method. Cox proportional hazards methods were used to estimate the effect of BMV and salvage modality on OS. RESULTS: Of 2829 patients, 2092 patients were included in the validation dataset. Of these, 921 (44.0%) experienced distant brain failure (DBF). Median OS from initial SRS was 11.2 mo. Median OS for BMV < 4, BMV 4-13, and BMV > 13 were 12.5 mo, 7.0 mo, and 4.6 mo (p < 0.0001). After multivariate regression modeling, melanoma histology (ß: 10.10, SE: 1.89, p < 0.0001) and number of initial brain metastases (ß: 1.52, SE: 0.34, p < 0.0001) remained predictive of BMV (adjusted R2 = 0.06). CONCLUSIONS: This multi-institutional dataset validates BMV as a predictor of OS following initial SRS. BMV is being utilized in upcoming multi-institutional randomized controlled trials as a stratification variable for salvage whole brain radiation versus salvage SRS after DBF.

CD138 plasma cells may predict brain metastasis recurrence following resection and stereotactic radiosurgery.

We sought to identify candidate biomarkers for early brain metastasis (BM) recurrence in patients who underwent craniotomy followed by adjuvant stereotactic radiosurgery. RNA sequencing was performed on eight resected brain metastasis tissue samples and revealed B-cell related genes to be highly expressed in patients who did not experience a distant brain failure and had prolonged overall survival. To translate the findings from RNA sequencing data, we performed immunohistochemistry to stain for B and T cell markers from formalin-fixed parffin-embedded tissue blocks on 13 patients. CD138 expressing plasma cells were identified and quantitatively assessed for each tumor sample. Patients' tumor tissues that expressed high levels of CD138 plasma cells (N = 4) had a statistically significant improvement in OS compared to low levels of CD138 (N = 9) (p = 0.01). Although these findings are preliminary, the significance of CD138 expressing plasma cells within BM specimens should be investigated in a larger cohort. Immunologic markers based on resection cavity analysis could be predictive for determining patient outcomes following cavity-directed SRS.

Cartilage oligomeric matrix protein in patients with osteoarthritis is independently associated with metastatic disease in prostate cancer.

Metastatic prostate cancer has a 5-year survival rate of 30%. Identifying predictors of metastasis outcome could potentially reduce patient mortality. The objective of this study was to determine whether osteoarthritis had an impact on outcomes of prostate cancer including death, local recurrence and/or metastasis and to determine whether cartilage oligomeric matrix protein was involved. We performed a retrospective case-control study of patients with prostate cancer with and without the diagnosis of osteoarthritis and completed immunohistochemistry (IHC) analysis of prostate (n=20) and lymph node (n=7) surgical specimens. We evaluated death, local recurrence and metastatic disease by various IHC biomarkers including prostate specific membrane antigen (PSMA), cartilage oligomeric matrix protein (COMP), CD31, and Ki-67. Our model identified osteoarthritis as an independent risk factor for metastatic disease (OR 5.24, 95% CI 1.49 - 18.41). Most notably, when joint arthroplasty was included in the model, osteoarthritis was no longer an independent risk factor for this outcome (p=0.071). IHC demonstrated that those with osteoarthritis, had greater expression of COMP in the prostate samples (mean 23.9% vs 5.84%, p<0.05) but not of Ki-67, CD31, or PSMA. This study identified and quantified increased metastatic disease in patients with osteoarthritis. Also, patients with osteoarthritis expressed increased COMP levels in the prostate and most likely in distant lymphatic nodes. Moreover, our findings suggest that joint arthroplasty may affect the ability of osteoarthritis to promote metastasis, which could impact treatment protocols and survival outcomes of the most common cause of cancer-related death (metastasis) in the United States.

Gamma Knife Radiosurgery for Multiple Sclerosis-Associated Trigeminal Neuralgia.

BACKGROUND: Trigeminal neuralgia in the setting of multiple sclerosis (MS-TN) is a challenging condition to manage that is commonly treated with Gamma Knife radiosurgery (GKRS; Elekta AB). However, data regarding the efficacy of this treatment are somewhat limited, particularly for repeat GKRS. OBJECTIVE: To report outcomes of GKRS for MS-TN from a cohort study. METHODS: Retrospective review of our GKRS database identified 77 cases of unilateral MS-TN (UMSTN) in 74 patients treated with GKRS between 2001 and 2016, with 37 cases undergoing repeat GKRS. Background medical history, treatment outcomes and complications, and dosimetric data were obtained by retrospective chart reviews and telephone interviews. RESULTS: Eighty-two percent of UMSTN cases achieved Barrow Neurological Institute (BNI) IIIb or better pain relief following initial GKRS for a median duration of 1.1 yr. Estimated rates of pain relief at 1, 3, and 5 yr were 51, 39, and 29% respectively. Eighty-eight percent achieved BNI IIIb or better pain relief after repeat GKRS for a median duration of 4.0 yr. Estimated rates of pain relief at 1 and 3 yr were 70 and 54%, respectively. Median doses for initial and repeat GKRS were 85 and 80 Gy to the 100% isodose line, respectively. Those with MS-TN had a shorter duration of BNI IIIb or better pain relief after initial (4.6 vs 1.1 yr), but not repeat GKRS (3.8 vs 4.0 yr) compared to a historical cohort from our institution. CONCLUSION: GKRS is an effective, well-tolerated treatment for patients with MS-TN. More durable relief is often achieved with repeat GKRS.

Initial SRS for Patients With 5 to 15 Brain Metastases: Results of a Multi-Institutional Experience.

PURPOSE: Several studies evaluating stereotactic radiosurgery (SRS) for patients with >4 brain metastases (BM) demonstrated similar outcomes after treatment of 1, 2 to 4, and 5 to 15 BM; others found clinically significant survival decrements in the latter group. In this review of 8 academic centers, we compared outcomes of patients undergoing initial SRS for 1, 2 to 4, and 5 to 15 BM. METHODS AND MATERIALS: A total of 2089 patients treated with initial SRS for BM were included. Overall survival (OS) was estimated using the Kaplan-Meier method and compared using the log-rank test. Patient and disease characteristics were evaluated for association with OS and cumulative incidence of distant brain failure (DBF) using stepwise multivariable Cox proportional hazards and competing risk regression modeling. RESULTS: In this series, 989 (47%) patients had 1 metastasis, 882 (42%) had 2 to 4 metastases, and 212 (10%) had 5 to 15 metastases treated. Median OS for the 1, 2 to 4, and 5 to 15 BM groups was 14.6, 9.5, and 7.5 months, respectively (log-rank P < .01). Univariate and multivariable analyses revealed no difference in survival between 2 to 4 and 5 to 15 BM. DBF at 1 year was 30%, 41%, and 50%, respectively (Gray's P < .01). Two-year cumulative incidence of salvage SRS decreased with increasing number of BM (1: 21% vs 2-4: 19% vs 5-15: 13%; P < .01), but no difference in salvage whole brain radiation therapy was observed (1: 12% vs 2-4: 15% vs 5-15: 16%, P = .10). At the time of DBF, median brain metastasis velocity was 3.9, 6.1, and 11.7 new metastases per year in the 1, 2 to 4, and 5 to 15 BM groups, respectively (P < .01). CONCLUSIONS: Patients treated with initial SRS for 5 to 15 BM experienced survival similar to that in patients with 2 to 4 BM. Lower rates of salvage SRS were observed in the 5 to 15 BM group, with no difference in rates of salvage whole brain radiation therapy.

Identification of CD37, cystatin A, and IL-23A gene expression in association with brain metastasis: analysis of a prospective trial.

PURPOSE/OBJECTIVES: We aimed to assess the predictive value of a lung cancer gene panel for the development of brain metastases. MATERIALS/METHODS: Between 2011 and 2015, 102 patients with lung cancer were prospectively enrolled in a clinical trial in which a diagnostic fine-needle aspirate was obtained. Gene expression was conducted on all samples that rendered a diagnosis of non-small cell lung cancer (NSCLC). Subsequent retrospective analysis of brain metastases-related outcomes was performed by reviewing patient electronic medical records. A competing risk multivariable regression was performed to estimate the adjusted hazard ratio for the development of brain metastases and non-brain metastases from NSCLC. RESULTS: A total of 49 of 102 patients had died by the last follow-up. Median time of follow-up was 13 months (range 0.23-67 months). A total of 17 patients developed brain metastases. Median survival time after diagnosis of brain metastases was 3.58 months (95% confidence interval (CI) 2.17, not available). A total of 30 patients developed metastases without any evidence of brain metastases until the time of death or last follow-up. Competing risk analysis identified three genes that were downregulated differentially in the patients with brain metastases versus non-brain metastatic disease: CD37 (0.017), cystatin A (0.022), and IL-23A (0.027). Other factors associated with brain metastases include: stage T ( P ⩽ 8.3e-6) and stage N ( P= 6.8e-4). CONCLUSIONS: We have identified three genes, CD37, cystatin A, and IL-23A, for which downregulation of gene expression was associated with a greater propensity for developing brain metastases. Validation of these biomarkers could have implications on surveillance patterns in patients with brain metastases from NSCLC.

Does Stereotactic Radiosurgery Have a Role in the Management of Patients Presenting With 4 or More Brain Metastases?

Stereotactic radiosurgery (SRS) and whole brain radiation therapy (WBRT) are effective treatments for management of brain metastases. Prospective trials comparing the 2 modalities in patients with fewer than 4 brain metastases demonstrate that overall survival (OS) is similar. Intracranial failure is more common after SRS, while WBRT is associated with neurocognitive decline. As technology has advanced, fewer technical obstacles remain for treating patients with 4 or more brain metastases with SRS, but level I data supporting its use are lacking. Observational prospective studies and retrospective series indicate that in patients with 4 or more brain metastases, performance status, total volume of intracranial disease, histology, and rate of development of new brain metastases predict outcomes more accurately than the number of brain metastases. It may be reasonable to initially offer SRS to some patients with 4 or more brain metastases. Initiating therapy with SRS avoids the acute and late sequelae of WBRT. Multiple phase III trials of SRS vs WBRT, both currently open or under development, are directly comparing quality of life and OS for patients with 4 or more brain metastases to help answer the question of SRS appropriateness for these patients.

Clinical Outcomes of Upfront Stereotactic Radiosurgery Alone for Patients With 5 to 15 Brain Metastases.

BACKGROUND: The role of primary stereotactic radiosurgery (SRS) for patients with >4 brain metastases (BM) remains controversial. OBJECTIVE: To compare the outcomes of patients treated with upfront SRS alone for 1, 2 to 4, and 5 to 15 BM and assess for predictors of clinical outcomes in the 5 to 15 BM group. METHODS: A total of 478 patients treated with upfront SRS were stratified by number of lesions: 220 had 1 BM, 190 had 2 to 4 BM, and 68 patients had 5 to 15 BM. Overall survival and whole brain radiotherapy-free survival were estimated using the Kaplan-Meier method. The cumulative incidences of local failure and distant brain failure (DBF) were estimated using competing risks methodology. Clinicopathologic and dosimetric parameters were evaluated as predictors of survival and DBF in patients with 5 to 15 BM using Cox proportional hazards. RESULTS: Median overall survival was 8.0, 6.3, and 4.7 mo for patients with 1, 2 to 4, and 5 to 15 BM, respectively (P = .14). One-year DBF was 27%, 44%, and 40%, respectively (P = .01). Salvage SRS and whole brain radiotherapy rates did not differ. Progressive extracranial disease and gastrointestinal primary were associated with poor survival while RCC primary was associated with increased risk of DBF. No evaluated dose-volume parameters predicted for death, neurologic death or toxicity. CONCLUSION: SRS for 5 to 15 BM is well tolerated without evidence of an associated increase in toxicity, treatment failure, or salvage therapy. Further prospective, randomized studies are warranted to clarify the role of SRS for these patients.

Analysis of the drivers of cost of management when patients with brain metastases are treated with upfront radiosurgery.

OBJECTIVES: We aimed to assess the driving factors for increased cost of brain metastasis management when using upfront stereotactic radiosurgery (SRS). PATIENT AND METHODS: 737 patients treated with upfront SRS without whole brain radiotherapy (WBRT). Patients were evaluated for use of craniotomy, length of hospital stay, need for rehabilitation or facility placement, and use of salvage SRS or salvage WBRT. Costs of care of these interventions were estimated based on 2013 Medicare reimbursements. Multiple linear regression was performed to determine factors that predicted for higher cost of treatment per month of life, as well as highest cumulative cost of care for brain metastasis. RESULTS: Mean cost of brain metastasis management per patient was $42,658, and $4673 per month of life. Upfront SRS represented the greatest contributor of total cost of brain metastasis management over a lifetime (49%), followed by use of any salvage SRS (21%), use of initial surgery (14%), use of salvage surgery (10%), hospitalization (3%) and cost of salvage WBRT (3%). Multiple linear regression identified brain metastasis velocity (BMV) (p < 0.001), use of cavity-directed SRS (<0.001), and CNS symptoms at time of presentation (p = 0.005) as factors that increased costs of care per month of survival. Use of salvage WBRT decreased per month cost of care in patients requiring salvage (p < 0.001). CONCLUSION: The cost of upfront SRS is the greatest contributor to cost of brain metastasis management when using upfront SRS. Higher BMV, progressive systemic disease and presence of symptoms are associated with increased cost of care.

Potential prognostic markers for survival and neurologic death in patients with breast cancer brain metastases who receive upfront SRS alone.

PURPOSE/OBJECTIVES: Stereotactic radiosurgery (SRS) is used as a treatment option for breast cancer brain metastases. It is unclear what factors predict neurologic death for these patients. MATERIALS/METHODS: A total of 128 patients with breast cancer brain metastases were treated with upfront SRS alone in this study. Survival was estimated using the Kaplan-Meier method. Clinicopathologic factors evaluated included age, ER/PR status, Her2 status, numbers of brain metastases treated, minimum SRS dose, disease-specific GPA, extracranial disease status and systemic disease burden. RESULTS: ER or PR positivity was associated with a trend towards decreased neurologic death (subdistribution hazard ratio (sHR) = 0.54, p=0.06). Factors associated with non-neurologic death include extracranial disease status (sHR = 2.02, p=0.02) and dose (sHR = 1.11, p=0.02); Her2-positivity was associated with reduced hazard of non-neurologic death (sHR 0.52, p=0.05). CONCLUSIONS: ER/PR positivity was associated with a trend towards less neurologic death. HER2 positivity was associated with a trend towards less non-neurologic death.

Impact of diabetes mellitus on outcomes in patients with brain metastasis treated with stereotactic radiosurgery.

PURPOSE: To determine the influence of diabetes mellitus (DM) on outcomes in patients with brain metastasis treated with stereotactic radiosurgery (SRS). METHODS: We retrospectively reviewed 498 patients with brain metastasis treated at our institution with SRS between January 2012 and March 2017. RESULTS: Eight-four patients (16.9%) held a diagnosis of DM prior to SRS treatment. Diabetics compared to nondiabetics had worse overall survival (OS). DM was found to be a significant predictor of OS on multivariate analysis (HR: 1.41, CI: 1.03-1.92, p = 0.03). When stratified by DM diagnosis, there were no significant differences in incidence of radiation necrosis (p = 0.82), radiation-induced edema (p = 0.88), cerebrospinal fluid leak (p = 0.49), or postoperative infection (p = 0.68). CONCLUSIONS: DM diagnosis was a significant predictor of poorer OS in patients treated for brain metastasis with SRS. Diabetics and nondiabetics experienced similar rates of radiation-associated brain toxicities.

Distinguishing True Progression From Radionecrosis After Stereotactic Radiation Therapy for Brain Metastases With Machine Learning and Radiomics.

PURPOSE: Treatment effect or radiation necrosis after stereotactic radiosurgery (SRS) for brain metastases is a common phenomenon often indistinguishable from true progression. Radiomics is an emerging field that promises to improve on conventional imaging. In this study, we sought to apply a radiomics-based prediction model to the problem of diagnosing treatment effect after SRS. METHODS AND MATERIALS: We included patients in the Johns Hopkins Health System who were treated with SRS for brain metastases who subsequently underwent resection for symptomatic growth. We also included cases of likely treatment effect in which lesions grew but subsequently regressed spontaneously. Lesions were segmented semiautomatically on preoperative T1 postcontrast and T2 fluid-attenuated inversion recovery magnetic resonance imaging, and radiomic features were extracted with software developed in-house. Top-performing features on univariate logistic regression were entered into a hybrid feature selection/classification model, IsoSVM, with parameter optimization and further feature selection performed using leave-one-out cross-validation. Final model performance was assessed by 10-fold cross-validation with 100 repeats. All cases were independently reviewed by a board-certified neuroradiologist for comparison. RESULTS: We identified 82 treated lesions across 66 patients, with 77 lesions having pathologic confirmation. There were 51 radiomic features extracted per segmented lesion on each magnetic resonance imaging sequence. An optimized IsoSVM classifier based on top-ranked radiomic features had sensitivity and specificity of 65.38% and 86.67%, respectively, with an area under the curve of 0.81 on leave-one-out cross-validation. Only 73% of cases were classifiable by the neuroradiologist, with a sensitivity of 97% and specificity of 19%. CONCLUSIONS: Radiomics holds promise for differentiating between treatment effect and true progression in brain metastases treated with SRS. A predictive model built on radiomic features from an institutional cohort performed well on cross-validation testing. These results warrant further validation in independent datasets. Such work could prove invaluable for guiding management of individual patients and assessing outcomes of novel interventions.

Glioblastoma radiomics: can genomic and molecular characteristics correlate with imaging response patterns?

PURPOSE: For glioblastoma (GBM), imaging response (IR) or pseudoprogression (PSP) is frequently observed after chemoradiation and may connote a favorable prognosis. With tumors categorized by the Cancer Genome Atlas Project (mesenchymal, classical, neural, and proneural) and by methylguanine-methyltransferase (MGMT) methylation status, we attempted to determine if certain genomic or molecular subtypes of GBM were specifically associated with IR or PSP. METHODS: Patients with GBM treated at two institutions were reviewed. Kaplan-Meier method was used to estimate overall survival (OS) and progression-free survival (PFS). Mantel-cox test determined effect of IR and PSP on OS and PFS. Fisher's exact test was utilized to correlate IR and PSP with genomic subtypes and MGMT status. RESULTS: Eighty-two patients with GBM were reviewed. The median OS and PFS were 17.9 months and 8.9 months. IR was observed in 28 (40%) and was associated with improved OS (median 29.4 vs 14.5 months p < 0.01) and PFS (median 17.7 vs 5.5 months, p < 0.01). PSP was observed in 14 (19.2%) and trended towards improved PFS (15.0 vs 7.7 months p = 0.08). Tumors with a proneural component had a higher rate of IR compared to those without a proneural component (IR 60% vs 28%; p = 0.03). MGMT methylation was associated with IR (58% vs 24%, p = 0.032), but not PSP (34%, p = 0.10). CONCLUSION: IR is associated with improved OS and PFS. The proneural subtype and MGMT methylated tumors had higher rates of IR.

Initial brain metastasis velocity: does the rate at which cancers first seed the brain affect outcomes?

PURPOSE/OBJECTIVE(S): Brain metastasis velocity (BMV) is a metric that describes the rate of development of new brain metastases (BM) after initial stereotactic radiosurgery (SRS). A limitation in the application of BMV is it cannot be applied until time of first BM failure after SRS. We developed initial BM velocity (iBMV), a new metric that accounts for the number of BM at first SRS and the time since initial cancer diagnosis. MATERIALS/METHODS: We reviewed patients with BM treated at our institution with upfront SRS without WBRT. iBMV was calculated as the number of BM at initial SRS divided by time (years) from initial cancer diagnosis to first SRS. We performed a linear regression to correlate BMV as a continuous variable and with low, intermediate, and high BMV risk groups. Kaplan-Meier estimation of OS was calculated from time of first SRS to death. iBMV was not calculated for patients who presented with BM at initial cancer diagnosis. RESULTS: 994 patients were treated with upfront SRS without WBRT between 2000 and 2017. Median OS was 8.5 mos. 595 (60%) patients developed BM after cancer diagnosis and median time to first SRS from time of initial diagnosis was 2.2 years. Median iBMV was 0.79 BM/year. iBMV correlated with BMV (ß = 1.57 p = 0.021) and independently predicted for mortality [Cox proportional hazard ratio (HR) 1.11, p = 0.036] after accounting for histology, number of initial brain metastases (HR 1.03, p = 0.32), time from cancer diagnosis to SRS (HR 0.98, p = 0.157) in a multivariate model. CONCLUSION: iBMV correlates with BMV and OS. With further validation, iBMV could serve as a metric to risk stratify patients for WBRT or SRS at time of first BM presentation.

Cortical Thinning and Structural Bone Changes in Non-Human Primates after Single-Fraction Whole-Chest Irradiation.

Stereotactic body radiation therapy (SBRT) is associated with an increased risk of vertebral compression fracture. While bone is typically considered radiation resistant, fractures frequently occur within the first year of SBRT. The goal of this work was to determine if rapid deterioration of bone occurs in vertebrae after irradiation. Sixteen male rhesus macaque non-human primates (NHPs) were analyzed after whole-chest irradiation to a midplane dose of 10 Gy. Ages at the time of exposure varied from 45-134 months. Computed tomography (CT) scans were taken 2 months prior to irradiation and 2, 4, 6 and 8 months postirradiation for all animals. Bone mineral density (BMD) and cortical thickness were calculated longitudinally for thoracic (T) 9, lumbar (L) 2 and L4 vertebral bodies; gross morphology and histopathology were assessed per vertebra. Greater mortality (related to pulmonary toxicity) was noted in NHPs <50 months at time of exposure versus NHPs >50 months ( P = 0.03). Animals older than 50 months at time of exposure lost cortical thickness in T9 by 2 months postirradiation ( P = 0.0009), which persisted to 8 months. In contrast, no loss of cortical thickness was observed in vertebrae out-of-field (L2 and L4). Loss of BMD was observed by 4 months postirradiation for T9, and 6 months postirradiation for L2 and L4 ( P < 0.01). For NHPs younger than 50 months at time of exposure, both cortical thickness and BMD decreased in T9, L2 and L4 by 2 months postirradiation ( P < 0.05). Regions that exhibited the greatest degree of cortical thinning as determined from CT scans also exhibited increased porosity histologically. Rapid loss of cortical thickness was observed after high-dose chest irradiation in NHPs. Younger age at time of exposure was associated with increased pneumonitis-related mortality, as well as greater loss of both BMD and cortical thickness at both in- and out-of-field vertebrae. Older NHPs exhibited rapid loss of BMD and cortical thickness from in-field vertebrae, but only loss of BMD in out-of-field vertebrae. Bone is sensitive to high-dose radiation, and rapid loss of bone structure and density increases the risk of fractures.