Predicting post-surgical functional status in high-grade glioma with resting state fMRI and machine learning - Scorecard - MDSpire

Predicting post-surgical functional status in high-grade glioma with resting state fMRI and machine learning

  • By

  • Patrick H. Luckett

  • Michael O. Olufawo

  • Ki Yun Park

  • Bidhan Lamichhane

  • Donna Dierker

  • Gabriel Trevino Verastegui

  • John J. Lee

  • Peter Yang

  • Albert Kim

  • Omar H. Butt

  • Milan G. Chheda

  • Abraham Z. Snyder

  • Joshua S. Shimony

  • Eric C. Leuthardt

  • May 24, 2024

  • 0 min

Share

Clinical Scorecard: Utilizing Resting State fMRI and Machine Learning to Forecast Postoperative Functional Outcomes in High-Grade Glioma Patients

At a Glance

CategoryDetail
ConditionHigh-grade gliomas (HGG), including glioblastoma multiforme (GBM)
Key MechanismsResting state functional MRI (RS-fMRI) assesses intrinsic brain network connectivity; machine learning models analyze RS-fMRI and clinical data to predict postoperative functional outcomes
Target PopulationAdult patients diagnosed with intracranial primary high-grade gliomas undergoing surgical resection
Care SettingNeurosurgical and neuro-oncology clinical settings with preoperative MRI imaging and postoperative functional assessment

Key Highlights

  • High-grade gliomas represent 60–70% of new brain tumor cases with median survival around 14 months despite standard treatment.
  • RS-fMRI enables functional brain network mapping without task performance, suitable even under sedation, with lower failure rates than task-based fMRI.
  • Machine learning models using RS-fMRI connectivity, tumor overlap with networks, tumor volume, and age can accurately predict postoperative functional outcomes (KPS ≥ 70 vs < 70).

Guideline-Based Recommendations

Diagnosis

  • Use histopathological and immunohistochemical criteria per WHO guidelines for definitive HGG diagnosis.
  • Acquire pre-surgical structural MRI (T1, T2, DTI) and resting state fMRI to evaluate tumor characteristics and brain functional networks.

Management

  • Perform maximal safe gross total resection followed by radiation and adjuvant chemoradiotherapy.
  • Incorporate RS-fMRI data into surgical planning to balance maximal tumor resection with functional preservation.

Monitoring & Follow-up

  • Assess postoperative functional status using Karnofsky Performance Scale (KPS) to evaluate patient outcomes.
  • Monitor genetic markers (MGMT methylation, EGFR amplification, TERT, IDH1, PTEN mutations) for prognostic insights.

Risks

  • Surgical resection may cause or worsen functional impairments impacting quality of life and overall survival.
  • Limitations of task-based fMRI include patient inability to perform tasks; RS-fMRI offers an alternative with fewer constraints.

Patient & Prescribing Data

Adults with newly diagnosed high-grade gliomas undergoing biopsy or surgical resection with available preoperative MRI data.

Machine learning models integrating RS-fMRI and clinical features can inform prognosis and guide surgical decision-making to optimize functional outcomes.

Clinical Best Practices

  • Utilize resting state fMRI for functional brain mapping preoperatively, especially in patients unable to perform task-based fMRI.
  • Apply machine learning techniques to extract relevant imaging and clinical features for individualized outcome prediction.
  • Balance extent of tumor resection with preservation of critical functional networks to improve postoperative quality of life.

References

Original Source(s)

Related Content