Methodological concerns on prostate‐specific membrane antigen ( PSMA) immunohistochemistry for predicting PSMA ‐ PET avidity

We read with great interest the study by Carll et al. [1] on the correlation between prostate-specific membrane antigen (PSMA) immunohistochemistry (IHC) staining and PSMA-positron emission tomography (PET) imaging findings. The study preliminarily explores the use of low-cost IHC technology to screen patients who may benefit from PSMA-PET examination, which holds significant value. Nevertheless, we would like to raise several critical methodological concerns. First, the study did not adjust for key clinicopathological confounders through multivariable analysis. The study employed univariate statistical methods to reveal the association between the PSMA IHC score and PET positivity concordance rate, but did not adjust for known confounders such as preoperative PSA, Prostate Imaging-Reporting and Data System (PI-RADS) score, and International Society of Urological Pathology (ISUP) Grade Group (GG), which are known to simultaneously affect both PSMA expression and PET uptake. Previous studies have confirmed that elevated ISUP GG is an independent predictor of high uptake on PSMA-PET [2]. Therefore, within the current analytical framework, it cannot be determined whether the predictive value of PSMA IHC is independent of tumour grade and other factors or merely serves as a surrogate marker for them. To establish the independent predictive efficacy of the IHC score, subsequent studies should employ multivariable logistic regression models for adjusted analyses. Second, the semi-quantitative visual scoring of IHC is subject to inter-observer variability, and the lack of a continuous variable correlation with the maximum standardised uptake value (SUVmax) limits predictive precision. The study employed a four-tier visual scoring system (0–3+), which is associated with substantial inter-observer variability, particularly in the classification of low-to-moderate expression levels where agreement is limited. More importantly, the study only analysed the binary PET concordance outcome and did not explore the dose–response relationship between IHC score and SUVmax, failing to provide a quantitative predictive basis for clinical practice. Adopting quantitative systems such as H-score or immunoreactive score, along with performing linear regression or Spearman correlation analysis, could more accurately reflect the quantitative association between PSMA expression and PET uptake, thereby enhancing its clinical applicability. Third, high PSMA expression in BPH tissue can cause false-positive staining, reducing the specificity of IHC assessment. The authors’ analysis assumes that IHC staining intensity reflects only PSMA expression in malignant tumours. However, Tang et al. [3] demonstrated that PSMA can also be highly expressed in atrophic epithelium and basal cells of BPH tissues, with a BPH-related false-positive rate of up to 30% on PSMA-PET. If the biopsy specimen contains such benign tissue with high PSMA expression, the IHC score will be artificially elevated, thereby overestimating its specificity for predicting PSMA-PET-positive results. In future study designs, the PSMA staining derived from malignant lesions vs benign background tissues should be carefully evaluated and distinguished. We commend the authors for their important work and look forward to future studies that address these methodological limitations to advance clinical practice. Yukun Liu: conceptualisation, writing – original draft. Zhangming Lei: writing – review and editing. Penglei Mao: supervision, writing – review and editing. None. The authors declare no conflicts of interest related to this work. This work was supported by the Quzhou Affiliated of Wenzhou Medical University, Quzhou People's Hospital. None.