LAUSR

We thank Dr. Campaña and Dr. Bernal for their interest in our recent study about the prospective comparison of 18F-PSMA-1007 PET/CT, whole-body MRI with diffusionweighted imaging (DWI) and contrast-enhanced CT in primary nodal staging of unfavourable intermediateand highrisk prostate cancer [1]. We appreciate their observations, and we would like to reply to their comments [2]. First of all, we thank the authors for pointing out the strength of 18F-PSMA-1007 radiotracer especially in the evaluation of the pelvic nodal status, due to its reduced urinary excretion. Despite the potential higher risk of false positive bone uptakes [3], 18F-PSMA-1007 appears to be an optimal tracer in the assessment of small lymph nodes close to the urinary tract. Besides, this is to our knowledge the first prospective study that evaluated this PSMA tracer in primary nodal staging of prostate cancer and compared its performance to whole-body MRI and contrast-enhanced CT. Surely, we agree with the authors that the relatively low percentage of patients with histopathological verification represents the main limitation of the study. However, the aim of this prospective trial was to assess the diagnostic performance of different imaging modalities in the overall primary staging (TNM) of prostate cancer. Results regarding M-staging have been published earlier [3], and T-staging results will be presented in a separate manuscript. Moreover, direct comparison of all three baseline imaging modalities (18F-PSMA-1007 PET/CT, WBMRI and CT) together with long imaging and clinical follow-up available for all patients allowed us to evaluate the nodal status with relatively good accuracy. Similar criteria have been used in the recent prospective randomized study by Hofman et al. [4], when histopathological verification was not available. Finally, the authors suggested using alternative methods to Cohen’s kappa coefficient to calculate the inter-reader agreement. Kendall’s coefficient is indeed preferably used when it comes to ordinal evaluations. However, our analysis was based on dichotomous variables (malignant/not malignant) in optimistic (equivocal lesions interpreted as benign) and pessimistic (equivocal lesions interpreted as malignant) analysis at the patient level. Therefore, the kappa coefficient seems to be the most appropriate method in this case [5], and, to our knowledge, there are no significant advantages in using Kendall’s over kappa with regard to dichotomous variables. Nevertheless, we have calculated Kendall’s coefficients at the patient level for the three imaging modalities, and our results are concordant with Kappa statistics, confirming the higher agreement between PSMA PET/CT readers (Table 1). This article is part of the Topical Collection on Oncology Genitourinary.