Stratification according to recursive partitioning analysis predicts outcome in newly diagnosed glioblastomas

Postoperative prognostic nomogram for adult grade II/III astrocytoma in the Chinese Han population

Background

Prognostic models of glioma have been the focus of many studies. However, most of them are based on Western populations. Additionally, because of the complexity of healthcare data in China, it is important to select a suitable model based on existing clinical data. This study aimed to develop and independently validate a nomogram for predicting the overall survival (OS) with newly diagnosed grade II/III astrocytoma after surgery.

Methods

Data of 472 patients with astrocytoma (grades II-III) were collected from Qilu Hospital as training cohort while data of 250 participants from Linyi People's Hospital were collected as validation cohort. Cox proportional hazards model was used to construct the nomogram and individually predicted 1-, 3-, and 5-year survival probabilities. Calibration ability, and discrimination ability were analyzed in both training and validation cohort.

Results

Overall survival was negatively associated with histopathology, age, subtotal resection, multiple tumors, lower KPS and midline tumors. Internal validation and external validation showed good discrimination (The C-index for 1-, 3-, and 5-year survival were 0.791, 0.748, 0.733 in internal validation and 0.754, 0.735, 0.730 in external validation, respectively). The calibration curves showed good agreement between the predicted and actual 1-, 3-, and 5-year OS rates.

Conclusion

This is the first nomogram study that integrates common clinicopathological factors to provide an individual probabilistic prognosis prediction for Chinese Han patients with astrocytoma (grades II-III). This model can serve as an easy-to-use tool to advise patients and establish optimized surveillance approaches after surgery.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13755-023-00223-0.

Prognostic factors of patients with Gliomas - an analysis on 335 patients with Glioblastoma and other forms of Gliomas

Background

The prognosis of glioma is poor, despite recent advances in diagnosis and treatment of the disease. It is important to investigate the clinical characteristics and prognostic factors of glioma so as to provide basis for treatment and management of patients.

Method

A total of 335 patients with glioma were included in this study. These patients were admitted to the medical center between November 2015 and December 2018. The clinical data, including demographic data, tumor characteristics, treatment strategy, expression pattern of tumor markers, and survival data, were retrospectively reviewed. Survival data were analyzed using Kaplan-Meier curves with log-rank test, while multivariate analysis Cox regression model was used to investigate risk factors for mortality.

Results

In this patient cohort, glioblastoma (40%), diffuse glioma (14.6%) and oligodendroglioma (9.6%) were the most common pathological types. The expression of Ki-67 was associated with several clinicopathological parameters (e.g. tumor type, grade, and number of lesions). In addition, Ki-67 correlated with the mortality within the first year of the post-treatment follow-up (P <  0.001). Kaplan-Maier analysis revealed that older patients (≥ 45 years) displayed worse prognosis than those aged under 45 years (P = 0.038). Dismal prognosis was also associated with clinical parameters, including high tumor grade, multiple lesions, and Karnofsky performance score (KPS). Multivariate analysis showed that low KPS (< 85) increased the risk of mortality by 2.3 folds with a 95% CI of 1.141 to 4.776 (P = 0.020). Low tumor grade (grade 1–2) oppositely reduced the mortality risk by 0.22 folds (95% CI, 0.065 to 0.763, P = 0.0168).

Conclusion

KPS and tumor grade were independent prognostic factors in patients with gliomas.

Amide proton transfer imaging might predict survival and IDH mutation status in high-grade glioma

OBJECTIVES

To assess the utility of amide proton transfer (APT) imaging as an imaging biomarker to predict prognosis and molecular marker status in high-grade glioma (HGG, WHO grade III/IV).

METHODS

We included 71 patients with pathologically diagnosed HGG who underwent preoperative MRI with APT imaging. Overall survival (OS) and progression-free survival (PFS) according to APT signal, clinical factors, MGMT methylation status, and IDH mutation status were analyzed. Multivariate Cox regression models with and without APT signal data were constructed. Model performance was compared using the integrated AUC (iAUC). Associations between APT signals and molecular markers were assessed using the Mann-Whitney test.

RESULTS

High APT signal was a significant predictor for poor OS (HR = 3.21, 95% CI = 1.62-6.34) and PFS (HR = 2.22, 95% CI = 1.33-3.72) on univariate analysis. On multivariate analysis, high APT signals were an independent predictor of poor OS and PFS when clinical factors alone (OS: HR = 2.89; PFS: HR = 2.13), or in combination with molecular markers (OS: HR = 2.85; PFS: HR = 2.00), were included as covariates. The incremental prognostic value of APT signals was significant for OS and PFS. IDH-wild type was significantly associated with high APT signals (p = 0.001) when compared to IDH-mutant; however, there was no difference based on MGMT methylation status (p = 0.208).

CONCLUSION

High APT signal was a significant predictor of poor prognosis in HGG. APT data showed significant incremental prognostic value over clinical prognostic factors and molecular markers and may also predict IDH mutation status.

KEY POINTS

• Amide proton transfer (APT) imaging is a promising prognostic marker of high-grade glioma. • APT signals were significantly higher in IDH-wild type compared to IDH-mutant high-grade glioma. • APT imaging may be valuable for preoperative screening and treatment guidance.

Relevance of a TCGA-derived Glioblastoma Subtype Gene-Classifier among Patient Populations

Glioblastoma multiforme (GBM), a deadly cancer, is the most lethal and common malignant brain tumor, and the leading cause of death in adult brain tumors. While genomic data continues to rocket, clinical application and translation to patient care are lagging behind. Big data now deposited in the TCGA network offers a window to generate novel clinical hypotheses. We hypothesized that a TCGA-derived gene-classifier can be applied across different gene profiling platforms and population groups. This gene-classifier validated three robust GBM-subtypes across six different platforms, among Caucasian, Korean and Chinese populations: Three Caucasian-predominant TCGA-cohorts (Affymetrix U133A = 548, Agilent Custom-Array = 588, RNA-seq = 168), and three Asian-cohorts (Affymetrix Human Gene 1.0ST-Array = 61, Illumina = 52, Agilent 4 × 44 K = 60). To understand subtype-relevance in patient therapy, we investigated retrospective TCGA patient clinical sets. Subtype-specific patient survival outcome was similarly poor and reflected the net result of a mixture of treatment regimens with/without surgical resection. As a proof-of-concept, in subtype-specific patient-derived orthotopic xenograft (PDOX) mice, Classical-subtype demonstrated no survival difference comparing radiation-therapy versus temozolomide monotherapies. Though preliminary, a PDOX model of Proneural/Neural-subtype demonstrated significantly improved survival with temozolomide compared to radiation-therapy. A larger scale study using this gene-classifier may be useful in clinical outcome prediction and patient selection for trials based on subtyping.

The Impact of Surgery on Long-Term Survival of Patients with Primary Gastric Diffuse Large B-Cell Lymphoma: A SEER Population-Based Study

BACKGROUND

The aim of this retrospective study was to compare the long-term survival of patients receiving conservative with surgical treatment to analyze the prognostic factors and the impact of surgery on oncological outcomes of patients with primary gastric diffuse large B-cell lymphoma.

METHODS

A total of 2647 patients diagnosed with primary gastric diffuse large B-cell lymphoma from 1998 to 2014 were extracted from SEER database. Propensity matching was performed to compare the clinicopathological characteristics of the two groups. Based on the recursive partitioning analysis, the patients were divided into three risk subgroups: low risk, intermediate risk, and high risk.

RESULTS

After propensity score matching, patient characteristics did not differ significantly between the two groups. The 5-year cancer-specific survival rates of the surgical group and the conservative treatment group were, respectively, 60% and 59.2% (P = 0.952) before propensity matching and 64.2% and 58.6% (P = 0.046) after propensity matching. According to the multivariate analysis, age, tumor stage, and chemotherapy and surgery were independent risk factors for long-term survival. The 5-year cancer-specific survival rates differed significantly between the low-risk, intermediate-risk, and high-risk patients (76.2% vs. 57.4% vs. 25.5%, respectively, P < 0.001). The 5-year cancer-specific survival rate of the surgical group was significantly higher than that of the conservative treatment group in the low-risk patients. However, it did not differ significantly in the intermediate-risk and high-risk patients (P > 0.05).

CONCLUSIONS

A prognostic model was constructed based on the independent risk factors of age, tumor stage, and chemotherapy. The prognostic model indicated that low-risk patients (age < 75 years, stage I/II, with/without chemotherapy) undergoing surgical treatment may benefit from long-term survival, while intermediate- and high-risk patients (age ≥ 75 years, stage I/II, with/without chemotherapy or III/IV patients, with/without chemotherapy) gain no significant benefit from surgery.

Negative Survival Impact of High Radiation Doses to Neural Stem Cells Niches in an IDH-Wild-Type Glioblastoma Population

Aims: Assess the impact of radiation doses to neural stem cell (NSC) niches in patients with IDH-wild-type glioblastoma.

Materials and Methods: Fifty patients were included in the study. NSC niches [SubVentricular Zone (SVZ) and Sub Granular Zone (SGZ)] were contoured by fusing CT scans and pre-therapy MRI, Tumor location defined ipsilateral and contralateral SVZ and SGZ. Prognostic significance of clinical, biological and dosimetric parameters were examined. We generated a Recursive Partitioning Analysis (RPA) model with independent prognostic classes.

Results: Median follow-up: 23.8 months. Event free and overall survival (OS): 10 and 19.1 months. Incomplete surgery, PTV (planning target volume), ipsilateral SVZ or NSC niche mean dose > 57.4 Gy, contralateral NSC niche mean dose > 35 Gy and bilateral NSC niche mean dose > 44 Gy were significantly correlated with reduced OS. Only EGFR amplification was an independent prognostic factor (p = 0.019) for OS. RPA generated independent risk groups: 1 (low risk): [ipsilateral NSC mean dose (INMD) < 58.01 Gy and methylated MGMT promoter], 2: (INMD < 58.01 Gy and unmethylated MGMT promoter and contralateral SVZ mean dose < 18.6 Gy; p = 0.43), 3: (INMD < 58.01 Gy and unmethylated MGMT promoter and contralateral SVZ mean dose > 18.6 Gy; p = 0.002) and 4: (very high risk) (INMD > 58.01 Gy; p < 0.001).

Conclusion: High radiation doses to ipsilateral NSC and contralateral SVZ could have a negative impact on overall survival in IDH-wild-type glioblastoma population.

A Comparative Analysis of the Usefulness of Survival Prediction Models for Patients with Glioblastoma in the Temozolomide Era: The Importance of Methylguanine Methyltransferase Promoter Methylation, Extent of Resection, and Subventricular Zone Location

OBJECTIVE

Several survival prediction models for patients with glioblastoma have been proposed, but none is widely used. This study aims to identify the predictors of overall survival (OS) and to conduct an independent comparative analysis of 5 prediction models.

METHODS

Multi-institutional data from 159 patients with newly diagnosed glioblastoma who received adjuvant temozolomide concomitant chemoradiotherapy (CCRT) were collected. OS was assessed by Cox proportional hazards regression and adjusted for known prognostic factors. An independent CCRT patient cohort was used to externally validate the 1) RTOG (Radiation Therapy Oncology Group) recursive partitioning analysis (RPA) model, 2) Yang RPA model, and 3) Wee RPA model, Chaichana model, and the RTOG nomogram model. The predictive accuracy for each model at 12-month survival was determined by concordance indices. Calibration plots were performed to ascertain model prediction precision.

RESULTS

The median OS for patients who received CCRT was 19.0 months compared with 12.7 months for those who did not (P < 0.001). Independent predictors were: 1) subventricular zone II tumors (hazard ratio [HR], 1.6; 95% confidence interval [CI], 1.0-2.5); 2) methylguanine methyltransferase promoter methylation (HR, 0.36; 95% CI, 0.2-0.6); and 3) extent of resection of >85% (HR, 0.59; 95% CI, 0.4-0.9). For 12-month OS prediction, the RTOG nomogram model was superior to the RPA models with a c-index of 0.70. Calibration plots for 12-month survival showed that none of the models was precise, but the RTOG nomogram performed relatively better.

CONCLUSIONS

The RTOG nomogram best predicted 12-month OS. Methylguanine methyltransferase promoter methylation status, subventricular zone tumor location, and volumetric extent of resection should be considered when constructing prediction models.