2027 - Diagnostic Performance of Treatment Response Assessment Maps In Detecting Tumor Recurrence vs. Adverse Radiation Effect after Brain Radiation

Purpose/Objective(s):

Differentiating tumor recurrence from adverse radiation effect (ARE) after brain radiation remains challenging. We evaluated Treatment Response Assessment Maps (TRAM), a delayed-contrast magnetic resonance imaging (MRI) technique that characterizes contrast washout patterns to distinguish persistent contrast extravasation (suggestive of ARE) from contrast clearance (suggestive of tumor recurrence), to determine its diagnostic performance in clinical use.

Materials/Methods:

Patients who underwent TRAM for worsening radiographic findings after radiotherapy for primary or metastatic central nervous system tumors, with or without magnetic resonance spectroscopy (MRS) and perfusion MRI (PERF), between May 2016 and February 2026 were included. Reference diagnosis was confirmed by at least one of the following: 1) pathology; 2) continued radiographic progression; or 3) progressive radiographic/clinical findings refractory to steroids. Diagnostic performance was assessed using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and overall accuracy. For combined analyses, recurrence was defined using parallel (OR logic: positive if either modality indicated recurrence) and serial (AND logic: positive if both modalities indicated recurrence) testing strategies.

Results:

A total of 58 patients with radiographic worsening in 75 brain lesions had evaluable TRAMs. Of these, 43 patients (59 lesions) had confirmed reference clinical diagnosis. Among the lesions, 28 were confirmed as tumor recurrence and 31 as ARE. TRAM alone demonstrated superior sensitivity of 75.0% and specificity of 93.5% for detecting tumor recurrence, compared with MRS alone (sensitivity 40.0%, specificity 88.2%), PERF alone (sensitivity 66.7%, specificity 82.6%) and the combined “MRS or PERF” (sensitivity 72.2%, specificity 70.6%). TRAM alone demonstrated the highest overall diagnostic accuracy (84.7%). Using a parallel testing strategy, the combined criterion of “TRAM or perfusion MRI” further improved sensitivity to 85.7%. Under a serial testing strategy, the combinations of “TRAM and MRS” and “TRAM and PERF” each achieved a specificity of 100%.

Conclusion:

TRAM alone demonstrated the highest accuracy in differentiating tumor recurrence from ARE after brain radiotherapy. Combining TRAM with PERF using a parallel testing strategy further enhanced sensitivity, while combining TRAM with either MRS or PERF under a serial testing approach improved specificity.

Modalities	Sensitivity	Specificity	PPV	NPV	Accuracy
TRAM	75.0%	93.5%	91.3%	80.6%	84.7%
MRS	40.0%	88.2%	80.0%	55.6%	62.2%
PERF	66.7%	82.6%	77.8%	73.1%	75.0%
TRAM or MRS	80.0%	88.2%	88.9%	78.9%	83.8%
TRAM or PERF	85.7%	82.6%	81.8%	86.4%	84.1%
MRS or PERF	72.2%	70.6%	72.2%	70.6%	71.4%
TRAM and MRS	35.0%	100.0%	100.0%	56.7%	64.9%
TRAM and PERF	57.1%	100.0%	100.0%	71.9%	79.5%
MRS and PERF	33.3%	94.1%	85.7%	57.1%	62.9%