Elizabeth R. Gerstner
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Elizabeth R. Gerstner has a total of 286 co-authors. Here are the top co-authors by number of shared publications:
- Jayashree Kalpathy-Cramer
Division of Artificial Intelligence Department of Ophthalmology University of Colorado Anschutz Medical Campus Aurora Colorado USA
25 shared publications - William T. Curry
Department of Neurosurgery, Massachusetts General Hospital, Boston, MA
24 shared publications - David A. Reardon
Harvard Medical School , Boston, Massachusetts
22 shared publications - Eudocia Quant Lee
Department of Neuro-Oncology, Dana-Farber Cancer Institute , Boston, MA
22 shared publications - Lakshmi Nayak
Harvard Medical School , Boston, Massachusetts
22 shared publications - Patrick Y. Wen
Harvard Medical School , Boston, Massachusetts
22 shared publications
Latest Publications
- 16 October 2026
- Published by Oxford University Press (OUP) in Neuro-Oncology
Prognostic significance of contrast enhancement in newly diagnosed grade 2 IDH-mutant glioma
Background: Contrast enhancement is generally considered an adverse prognostic feature in adult glioma. However, its significance in newly diagnosed grade 2 isocitrate dehydrogenase (IDH)-mutant glioma remains uncertain, with important implications for use of the mutant IDH inhibitor vorasidenib. We evaluated survival outcomes associated with preoperative enhancement in this population. Methods: We retrospectively evaluated consecutive patients with grade 2 IDH-mutant glioma treated at 2 institutions between 2010 and 2021. Preoperative MRIs were assessed for intratumoral enhancement by 2 independent readers with third-reader adjudication. Overall survival (OS) and progression-free survival (PFS) were evaluated by Kaplan-Meier analysis and univariate and multivariable Cox regression. Results: Among 325 evaluable patients, the median age was 37 years, 47% had 1p/19q-codeletion, and 35% had gross total resection. Median follow-up was 10.5 years. Preoperative enhancement was observed in 53 (16%) patients, most commonly nonmeasurable (72%) or ill-defined (62%); most (87%) had no residual postoperative enhancement. Preoperative enhancement was not associated with OS or PFS in either univariate analyses (OS hazard ratio [HR] 1.16 [0.67-2.01], P = .591; PFS HR 1.25 [0.88-1.80], P = .216) or multivariable analyses adjusting for established prognostic factors (OS HR 0.86 [0.46-1.61], P = .633; PFS HR 1.29 [0.86-1.92], P = .219). Enhancement characteristics (measurable size, well-defined morphology) were not associated with OS. Conclusions: In grade 2 IDH-mutant glioma, preoperative enhancement was not associated with OS or PFS. These findings suggest that preoperative enhancement does not identify a higher-risk subgroup in this population and that its presence alone should not preclude consideration of vorasidenib in eligible patients.
- 16 October 2026
- Published by Oxford University Press (OUP) in Neuro-Oncology
Prognostic significance of contrast enhancement in newly diagnosed grade 2 IDH-mutant glioma
Background: Contrast enhancement is generally considered an adverse prognostic feature in adult glioma. However, its significance in newly diagnosed grade 2 isocitrate dehydrogenase (IDH)-mutant glioma remains uncertain, with important implications for use of the mutant IDH inhibitor vorasidenib. We evaluated survival outcomes associated with preoperative enhancement in this population. Methods: We retrospectively evaluated consecutive patients with grade 2 IDH-mutant glioma treated at 2 institutions between 2010 and 2021. Preoperative MRIs were assessed for intratumoral enhancement by 2 independent readers with third-reader adjudication. Overall survival (OS) and progression-free survival (PFS) were evaluated by Kaplan-Meier analysis and univariate and multivariable Cox regression. Results: Among 325 evaluable patients, the median age was 37 years, 47% had 1p/19q-codeletion, and 35% had gross total resection. Median follow-up was 10.5 years. Preoperative enhancement was observed in 53 (16%) patients, most commonly nonmeasurable (72%) or ill-defined (62%); most (87%) had no residual postoperative enhancement. Preoperative enhancement was not associated with OS or PFS in either univariate analyses (OS hazard ratio [HR] 1.16 [0.67-2.01], P = .591; PFS HR 1.25 [0.88-1.80], P = .216) or multivariable analyses adjusting for established prognostic factors (OS HR 0.86 [0.46-1.61], P = .633; PFS HR 1.29 [0.86-1.92], P = .219). Enhancement characteristics (measurable size, well-defined morphology) were not associated with OS. Conclusions: In grade 2 IDH-mutant glioma, preoperative enhancement was not associated with OS or PFS. These findings suggest that preoperative enhancement does not identify a higher-risk subgroup in this population and that its presence alone should not preclude consideration of vorasidenib in eligible patients.
- Open Access
- 12 January 2026
- Published by Oxford University Press (OUP) in Neuro-Oncology Advances
- Vol. 8 (1)
A deep learning algorithm for fully automated volumetric measurement of meningioma burden
Background: We sought to develop a deep learning (DL) model to enable fully automated 3D segmentation and volumetric assessment of meningioma burden with a specific emphasis on generalizing to high-grade and posttreatment meningiomas to improve interobserver variability and decrease reader time investment in tumor response assessment. Methods: In total, 450 postcontrast T1-weighted brain MRIs from 104 patients with meningiomas were obtained from Massachusetts General Hospital and Dana-Farber Cancer Institute. The cohort was unique among prior DL segmentation models in that it encompassed meningiomas of all grades, postoperative, and postradiated meningiomas. Preprocessed MRIs and manually generated tumor segmentations were used to train a U-Net with a joint Dice-cross entropy loss function. Results: When tested on internal data, our model achieved a median Dice of 0.741 and a median 95th percentile Hausdorff Distance (HD95) of 26 mm on a high-grade test set and a median Dice of 0.848 and median HD95 of 1.41 mm on a test set with low-grade tumors. Lesion-wise metrics were equivalent to global metrics for low-grade tumors, which contained only single lesions, but were substantially lower for high-grade tumors, with a median lesion-wise Dice of 0.45 and median lesion-wise HD95 of 130 mm, reflecting greater difficulty delineating individual high-grade lesions. Our model also generalized well to 1000 studies from 944 patients selected from the public BraTS dataset, achieving a median Dice of 0.923 and median HD95 of 2.24 mm. Conclusions: The study produced a model that addresses an unmet need for automated volumetric measurements of meningiomas and created a reliable metric for quantifying meningioma burden. In comparison to prior DL approaches, our model achieved competitive performance on external data and improved Dice scores on high-grade and posttreatment meningiomas. The trained model, volumetric evaluation code, and accompanying documentation are available online at https://github.com/mccle/tumor_segmentation.
- 12 January 2026
- Published by Oxford University Press (OUP) in Neuro-Oncology Advances
- Vol. 8 (1)
Open AccessA deep learning algorithm for fully automated volumetric measurement of meningioma burden
Background: We sought to develop a deep learning (DL) model to enable fully automated 3D segmentation and volumetric assessment of meningioma burden with a specific emphasis on generalizing to high-grade and posttreatment meningiomas to improve interobserver variability and decrease reader time investment in tumor response assessment. Methods: In total, 450 postcontrast T1-weighted brain MRIs from 104 patients with meningiomas were obtained from Massachusetts General Hospital and Dana-Farber Cancer Institute. The cohort was unique among prior DL segmentation models in that it encompassed meningiomas of all grades, postoperative, and postradiated meningiomas. Preprocessed MRIs and manually generated tumor segmentations were used to train a U-Net with a joint Dice-cross entropy loss function. Results: When tested on internal data, our model achieved a median Dice of 0.741 and a median 95th percentile Hausdorff Distance (HD95) of 26 mm on a high-grade test set and a median Dice of 0.848 and median HD95 of 1.41 mm on a test set with low-grade tumors. Lesion-wise metrics were equivalent to global metrics for low-grade tumors, which contained only single lesions, but were substantially lower for high-grade tumors, with a median lesion-wise Dice of 0.45 and median lesion-wise HD95 of 130 mm, reflecting greater difficulty delineating individual high-grade lesions. Our model also generalized well to 1000 studies from 944 patients selected from the public BraTS dataset, achieving a median Dice of 0.923 and median HD95 of 2.24 mm. Conclusions: The study produced a model that addresses an unmet need for automated volumetric measurements of meningiomas and created a reliable metric for quantifying meningioma burden. In comparison to prior DL approaches, our model achieved competitive performance on external data and improved Dice scores on high-grade and posttreatment meningiomas. The trained model, volumetric evaluation code, and accompanying documentation are available online at https://github.com/mccle/tumor_segmentation.
- 15 December 2025
- Published by Oxford University Press (OUP) in Neuro-Oncology
- Vol. 28 (3) , 815-818
Letter to the EditorTumor-infiltrating lymphocyte therapy generated from melanoma brain metastasis: Proof-of-concept for CNS-derived cellular immunotherapy
- 15 December 2025
- Published by Oxford University Press (OUP) in Neuro-Oncology
- Vol. 28 (3) , 815-818
Letter to the EditorTumor-infiltrating lymphocyte therapy generated from melanoma brain metastasis: Proof-of-concept for CNS-derived cellular immunotherapy
- 13 May 2025
- Published by Oxford University Press (OUP) in Neuro-Oncology Practice
- Vol. 12 (5) , 797-802
A single-institution study of the natural history of fibroblast growth factor receptor-altered gliomas
Background: Fibroblast growth factor receptor (FGFR) alterations are potential oncogenic drivers that occur in a subset of patients with gliomas, but the natural history of these tumors is not clearly defined. An understanding of outcomes of FGFR-driven glioma has implications for targeted drug development for FGFR inhibitors, which are approved for other cancers. Methods: We performed a retrospective cohort study of patients with gliomas who harbored FGFR alterations, including fusions, single nucleotide variant (SNV), and copy number variant (CNV) alterations seen at Massachusetts General Hospital between 2003 and 2023. The electronic medical record was searched to identify additional molecular data, treatment, and MRI scans. Kaplan–Meier analysis was used to assess progression-free and overall survival (OS). Results: Thirty-one patients with glioblastoma (GBM), diffuse astrocytoma, and glioneuronal tumors with FGFR alterations were identified: 17 with FGFR fusions, 10 with SNV, and 4 with CNV. FGFR3-TACC3 was the most common fusion in patients with GBM or diffuse astrocytoma. Median OS in patients with GBM was 2.75 years, despite 55% of tumors having an unmethylated MGMT promoter. There were no clearly co-occurring mutations with an FGFR alteration. Of 7 patients who underwent subsequent surgery, 6 lost the original FGFR alteration. No patient received an FGFR inhibitor. Conclusions: While FGFR alterations are rare in glioma, patients with FGFR-altered GBM may have prolonged survival, which has implications for clinical trial design. We found loss of FGFR alteration at the time of subsequent surgery, raising concern for the therapeutic potential of FGFR-targeting agents in recurrent gliomas.
- 13 May 2025
- Published by Oxford University Press (OUP) in Neuro-Oncology Practice
- Vol. 12 (5) , 797-802
A single-institution study of the natural history of fibroblast growth factor receptor-altered gliomas
Background: Fibroblast growth factor receptor (FGFR) alterations are potential oncogenic drivers that occur in a subset of patients with gliomas, but the natural history of these tumors is not clearly defined. An understanding of outcomes of FGFR-driven glioma has implications for targeted drug development for FGFR inhibitors, which are approved for other cancers. Methods: We performed a retrospective cohort study of patients with gliomas who harbored FGFR alterations, including fusions, single nucleotide variant (SNV), and copy number variant (CNV) alterations seen at Massachusetts General Hospital between 2003 and 2023. The electronic medical record was searched to identify additional molecular data, treatment, and MRI scans. Kaplan–Meier analysis was used to assess progression-free and overall survival (OS). Results: Thirty-one patients with glioblastoma (GBM), diffuse astrocytoma, and glioneuronal tumors with FGFR alterations were identified: 17 with FGFR fusions, 10 with SNV, and 4 with CNV. FGFR3-TACC3 was the most common fusion in patients with GBM or diffuse astrocytoma. Median OS in patients with GBM was 2.75 years, despite 55% of tumors having an unmethylated MGMT promoter. There were no clearly co-occurring mutations with an FGFR alteration. Of 7 patients who underwent subsequent surgery, 6 lost the original FGFR alteration. No patient received an FGFR inhibitor. Conclusions: While FGFR alterations are rare in glioma, patients with FGFR-altered GBM may have prolonged survival, which has implications for clinical trial design. We found loss of FGFR alteration at the time of subsequent surgery, raising concern for the therapeutic potential of FGFR-targeting agents in recurrent gliomas.
- Open Access
- 1 April 2024
- Published by American Association for Cancer Research (AACR)
Supplementary Table S5 from Clinical and Genomic Predictors of Adverse Events in Newly Diagnosed Glioblastoma
Supplemental Table 5: Association between genetic alterations and age at diagnosis showing significant genes at a false discovery rate (FDR)<0.2.
- 1 April 2024
- Published by American Association for Cancer Research (AACR)
Open AccessSupplementary Table S5 from Clinical and Genomic Predictors of Adverse Events in Newly Diagnosed Glioblastoma
Supplemental Table 5: Association between genetic alterations and age at diagnosis showing significant genes at a false discovery rate (FDR)<0.2.
- Open Access
- 1 April 2024
- Published by American Association for Cancer Research (AACR)
Supplementary Table S3 from Clinical and Genomic Predictors of Adverse Events in Newly Diagnosed Glioblastoma
Supplemental Table 3A: Univariate analysis of molecular alterations associated with seizure at diagnosis showing significant genes at a nominal p<0.05. Genes associated with seizures at a false discovery rate (FDR) <0.2 were tested in multivariable analyses (see Supplemental Table 4).
- 1 April 2024
- Published by American Association for Cancer Research (AACR)
Open AccessSupplementary Table S3 from Clinical and Genomic Predictors of Adverse Events in Newly Diagnosed Glioblastoma
Supplemental Table 3A: Univariate analysis of molecular alterations associated with seizure at diagnosis showing significant genes at a nominal p<0.05. Genes associated with seizures at a false discovery rate (FDR) <0.2 were tested in multivariable analyses (see Supplemental Table 4).
- Open Access
- 1 April 2024
- Published by American Association for Cancer Research (AACR)
Supplementary Table S4 from Clinical and Genomic Predictors of Adverse Events in Newly Diagnosed Glioblastoma
Supplemental Table 4: Multivariable analysis of molecular alterations associated with seizure at presentation showing significant genes at a false discovery rate (FDR) <0.2. Multivariable models included preoperative tumor volume, tumor location, and age at diagnosis. Abbreviations: GoF, gain of function; LoF, loss of function.
- 1 April 2024
- Published by American Association for Cancer Research (AACR)
Open AccessSupplementary Table S4 from Clinical and Genomic Predictors of Adverse Events in Newly Diagnosed Glioblastoma
Supplemental Table 4: Multivariable analysis of molecular alterations associated with seizure at presentation showing significant genes at a false discovery rate (FDR) <0.2. Multivariable models included preoperative tumor volume, tumor location, and age at diagnosis. Abbreviations: GoF, gain of function; LoF, loss of function.