Tozuleristide Fluorescence-Guided Resection of Gliomas: Results of a Phase II Investigator-Initiated Study

This Phase II, investigator-initiated, nonrandomized clinical trial evaluated tozuleristide, a tumor-specific near-infrared fluorescence imaging agent, combined with the Canvas® imaging system (Blaze Bioscience) to enhance intraoperative visualization of glial tumors.

Subjects with newly diagnosed, suspected, biopsy-proven, or previously resected high- or low-grade central nervous system tumors received 24 mg (contrast-enhancing glioma) or 36 mg (non-contrast-enhancing glioma) of tozuleristide 1 to 24 hours before surgery. Tumors were visualized intraoperatively using the Canvas imaging system. Biopsies were collected to assess tozuleristide sensitivity and specificity by comparing with histopathology assessment. Extent of resection (EOR), calculated from preoperative and postoperative MRI, was compared with historical controls. Fluorescence did not guide resection, and subjects were not selected for maximal EOR.

Thirty-nine subjects (33 contrast-enhancing, 6 nonenhancing tumors) underwent all study procedures. Fluorescence was observed in 30/33 contrast-enhancing tumors. Across 191 biopsies, fluorescence showed sensitivity = 0.735, specificity = 0.500, positive predictive value (PPV) = 0.847, negative predictive value = 0.333, and accuracy = 0.685. Residual fluorescence occurred only when total resection was unsafe and showed a nonsignificant negative correlation with EOR (r = -0.318). The mean EOR was 90.09% (24 mg tozuleristide) with 12 subjects achieving gross-total resection; 0/6 nonenhancing tumors fluoresced. Among patients with glioblastoma undergoing repeat resection, exploratory comparison with historical controls showed higher gross-total resection rates in tozuleristide-treated subjects (6/11 vs 2/11, P = .076); however, fluorescence was not incorporated into surgical decision-making, and these results thus cannot be interpreted toward efficacy.

Tozuleristide demonstrated tumor-specific fluorescence with high sensitivity and PPV at 24 mg. The high PPV suggests that fluorescence may be useful for confirming tumor presence in enhancing gliomas, but the low negative predictive value seen cannot exclude tumor in the absence of fluorescence, particularly in infiltrative margins. A well-controlled Phase III trial is necessary to determine the efficacy of image-guided resection in improving EOR.