A histogram of [¹⁸F]BBPA PET imaging differentiates non-neoplastic lesions from malignant brain tumors

Differentiating treatment response from tumor progression is fundamental but challenging for almost all oncological subjects, as the treatment strategy is effective and should be insisted in the former situation, while the therapeutic regimen is invalid and necessitates substitutions in the latter circumstances [1]. However, radiotherapy or immunotherapy may induce pseudo-progression, a transient increase of tumor volume due to tumor cell lysis or immune cell infiltration followed by delayed tumor shrinkage, and is difficult for early clinical and radiological identification [1, 2]. In malignant brain tumors, 10–30% of tumors showed pseudo-progression following radiotherapy, immunotherapy and targeted therapy [3‐6], some of which were not restricted to the recent onset of treatment [7, 8]. In addition, alternative non-neoplastic conditions such as radiation necrosis or inflammation may also mimic neoplasms and warrant appropriate distinction [3]. Response Evaluation Criteria in Solid Tumors (RECIST) and Response Assessment in Neuro­Oncology (RANO) have been proposed [9, 10], yet the performance in distinguishing treatment response from tumor progression remains to be improved [11‐14].

Boramino acids (BAA) are a class of amino acid biosimilars with the boron trifluoride group (–BF₃) to replace the carboxyl group (–COOH) of amino acids, which mimics the corresponding amino acid in biological recognition and transportation [15]. The ¹⁸F-¹⁹F isotope exchange reaction of boron trifluoride moiety allows the molecule to be mildly radiolabeled and can facilitate tumor theranostics through identical chemical structure (the only difference between positron emission tomography [PET] diagnosis and boron neutron capture therapy [BNCT] for treatment is ¹⁸F or ¹⁹F) [15‐20]. The first-in-human study of this class of PET tracers demonstrated sufficient safety, clean background and high tumor activity in malignant brain tumors [21, 22], validating the concept and potential clinical value of boron amino acids. Subsequently, trifluoroborate boronophenylalanine (BBPA) that replaced the carboxyl group (–COOH) of 4‑boronophenylalanine (BPA) with boron trifluoride group (-BF₃) was synthesized and is recognized as the next generation of boron amino acids thanks to the doubled boron delivery efficiency [23].

This study raised a [¹⁸F]BBPA PET-based approach to differentiate non-neoplastic lesions from proliferating tumors, aiming to provide a non-invasive method to uncover true lesion property. A total of 21 patients were included and underwent [¹⁸F]BBPA PET and contrast-enhanced magnetic resonance imaging (MRI) scans. Both neoplastic and non-neoplastic lesions exhibited elevated [¹⁸F]BBPA radioactivity and cannot be distinguished by traditional parameters. Histograms of the standard uptake value (SUV) within region of interest (ROI) were plotted, and the malignant tumors exhibited a symmetrical distribution (similar to normal distribution), while the non-neoplastic lesions displayed a positive skewed (left deviated) distribution. Such difference can be further quantified by skewness and tendency, providing an alternative method for differential diagnosis.

Differentiating neoplastic and non-neoplastic lesions (i.e., inflammation, necrosis, anti-tumor immune response) remains a critical clinical issue at both initial diagnosis and treatment follow-up. Amino acid tracers such as [¹⁸F]FET were investigated to distinguish tumor progression and treatment-related changes, with a T/N ratio displayed accuracy of 0.70 and area under the ROC curve (AUC) of 0.75 at a cutoff value of 1.95 [26]. However, considerable situations were not identified by traditional parameters, and both neoplastic and non-neoplastic lesions exhibited elevated [¹⁸F]BBPA activity. Histogram was further proposed for differential diagnosis, and the SUV of a normal or neoplastic area with regional heterogeneity (e.g., [¹⁸F]FDG in the brain, [¹⁸F]FDG or [¹⁸F]FLT in head and neck squamous cell carcinoma) are expected to be normal distribution [27, 28]. The non-neoplastic lesions displayed positively skewed (left deviated) voxel value distribution that was visually differed from the normally distributed neoplastic lesions on the histogram, and can be further quantified by skewness and tendency, providing an alternative method for differential diagnosis. The clinical impact is further demonstrated in recent cases, in which [¹⁸F]BBPA PET identified the lesion properties earlier than traditional methods. Therefore, the histogram of [¹⁸F]BBPA PET might aid the differentiation of neoplastic and non-neoplastic lesions and ultimately facilitate the accurate treatment decisions.

The histogram analysis may be applied to other circumstances (i.e., other disease or radiotracers) with low background activity and high lesion uptake, and the segmentation is preferably conducted on alternative imaging modality rather than PET image (threshold-based PET segmentation would result in a clear boundary on histogram). However, the current study had several limitations including a small sample size (particularly for non-neoplastic lesions) and a short follow-up period (unable to demonstrate the prognostic value of [¹⁸F]BBPA histogram). For future works, a well-designed prospective study with balanced cohort and longitudinal follow-up is necessary to validate the findings, and an in-depth exploration of the mechanism underlying the [¹⁸F]BBPA histogram differences is necessitated. In conclusion, the histogram of [¹⁸F]BBPA PET can differentiate non-neoplastic lesions from proliferating tumors and would facilitate the precision diagnosis and patient management.