Central Nervous System

PV QA 2 - Poster Viewing Q&A 2

MO_6_2622 - Validation and optimization of a web-based nomogram for predicting survival of patients with newly diagnosed glioblastoma

Monday, October 22
10:45 AM - 12:15 PM
Location: Innovation Hub, Exhibit Hall 3

Validation and optimization of a web-based nomogram for predicting survival of patients with newly diagnosed glioblastoma
N. Kim1, J. Chang1, C. W. Wee2, I. A. Kim2, J. H. Chang3, H. S. Lee4, J. H. Cho1, S. H. Kim5, S. G. Kang3, J. W. Kim6, C. K. Hong7, E. Kim2, T. M. Kim8, Y. J. Kim8, C. K. Park9, J. W. Kim9, C. Kim9, S. H. Choi10, J. H. Kim10, S. H. Park11, G. Choe11, S. T. Lee12, I. H. Kim2, and C. O. Suh1; 1Department of Radiation Oncology, Yonsei Cancer Center, Yonsei University College of Medicine, Seoul, Korea, Republic of (South), 2Department of Radiation Oncology, Seoul National University College of Medicine, Seoul, Korea, Republic of (South), 3Department of Neurosurgery, Yonsei Cancer Center, Yonsei University College of Medicine, Seoul, Korea, Republic of (South), 4Department of Biostatistics, Yonsei University College of Medicine, Seoul, Korea, Republic of (South), 5Department of Pathology, Yonsei Cancer Center, Yonsei University College of Medicine, Seoul, Korea, Republic of (South), 6Department of Radiation Oncology, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Korea, Republic of (South), 7Department of Neurosurgery, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Korea, Republic of (South), 8Department of Internal Medicine, Seoul National University College of Medicine, Seoul, Korea, Republic of (South), 9Department of Neurosurgery, Seoul National University College of Medicine, Seoul, Korea, Republic of (South), 10Department of Radiology, Seoul National University College of Medicine, Seoul, Korea, Republic of (South), 11Department of Pathology, Seoul National University College of Medicine, Seoul, Korea, Republic of (South), 12Department of Neurology, Seoul National University College of Medicine, Seoul, Korea, Republic of (South)

Purpose/Objective(s): The extent of resection (EOR) was recognized as an important prognostic factor of glioblastoma in the temozolomide (TMZ) era. However, the EOR was determined based on the neurosurgeon’s judgment in both the 2008 EORTC-NCIC nomogram and 2016 NRG nomogram, which shows greater relative subjectivity in the assessment of the EOR than postoperative radiological confirmation. This collaborative research was undertaken to assess the validity of the recent NRG nomogram and further optimize the nomogram by using a multicenter dataset where postoperative MRI evaluation is commonly used in daily practice.

Materials/Methods: We identified patients with newly diagnosed glioblastoma, treated with a TMZ-based radio-chemotherapy between 2006 and 2016 at four hospitals. Patients who had unknown O6-methylguanin-DNA-methyltransferase (MGMT) status, secondary glioblastoma, leptomeningeal seeding, and no perioperative brain MRI were excluded. While the NRG nomogram used the surgeon-defined EOR, the MRI-defined EOR was used in this study. The validity of the prediction algorithm was assessed regarding discrimination and calibration. If additional available factors had prognostic value independent of the NRG nomogram, models were developed using data from three hospitals and were independently validated using data from another hospital. Models were internally validated using cross-validation and bootstrapping. Results: A total of 837 patients met the eligibility criteria. The median overall survival (OS) was 21.1 months. The NRG nomogram severely underestimated the 6-, 12-, and 24-month survival probabilities in all the patients. The overall discrepancy was greater in predicting long-term survival probabilities (Harrell’s c-index at 12 and 24 months: 0.618 and 0.606, respectively) than in predicting 6-month survival probability (c-index: 0.700). In the multivariable analysis, unlike the original NRG nomogram, MRI-defined total resection had a higher impact on OS, and two additional factors, namely the IDH1 mutation and tumor contacting subventricular zone, had independent prognostic values. The updated nomogram was built using the Cox proportional hazards model with these variables. After internal validation, c-index was calculated for predicting survival at 6, 12, and 24 months. In all the three time-points, the updated nomogram outperformed the NRG nomogram (6-, 12-, and 24-month: 0.728, 0.688, and 0.688, respectively) and was well calibrated. The updated nomogram was also validated using an independent dataset. For all the three time-points, the updated nomogram was well calibrated and accurately predicted 6-, 12-, and 24-month survivals (c-index for 6, 12, and 24 months: 0.787, 0.751, and 0.719, respectively). Conclusion: The performance of the NRG nomogram was suboptimal, with underestimation of the 12- and 24-month survival probabilities. The updated nomogram for assessing OS estimates for patients with glioblastoma has been optimized and validated.

Author Disclosure: N. Kim: None. J. Chang: None. S. Kang: None.

Send Email for Nalee Kim


Assets

MO_6_2622 - Validation and optimization of a web-based nomogram for predicting survival of patients with newly diagnosed glioblastoma



Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for Validation and optimization of a web-based nomogram for predicting survival of patients with newly diagnosed glioblastoma