This study was approved by the ethics committee of Fujian Provincial Hospital (reference number, K2019-10-017) and conducted in compliance with the principles of the Declaration of Helsinki. Furthermore, informed consent was obtained from all participants and/or their legal guardians.

We conducted a prospective head-to-head observational study to analyze the diagnostic efficacy between ^99mTc-PSMA SPECT/CT and mpMRI in treatment-naive PCa. In this study, the sensitivity and specificity of ^99mTc-PSMA SPECT/CT and mpMRI in the diagnosis of PCa were assumed to be greater than 50% (H_0 = 50%). Referring to similar published literature on ⁶⁸Ga-PSMA PET/CT and mpMRI (11), a sensitivity and specificity value of 80% was assumed. PASS 11 software (Power Analysis and Sample Size, NCSS, LLC) was used to estimate the required sample size. Assuming α=0.05 (unilateral), β=0.1, and a 1:1 ratio between the groups, the calculations indicated that at least 46 patients needed to be included in the study. Consequently, 56 individuals were enrolled in this study.

Fifty-six men were enrolled in this study between October 2019 and November 2022. The inclusion criteria were as follows (2): ① digital rectal examination touching the prostate nodules; ② transrectal ultrasound suspected PCa; ③ PSA>10 ng/mL or progressive PSA increase (12); ④ no treatment administered before the scan; and ⑤ complete medical records, control data, and clinical follow-up results. The exclusion criteria were as follows: ① the presence of severe syndromes that were difficult to manage; ② active or upcoming participation in other clinical drug trials; ③ lack of regular review or follow-up results; ④ a second primary tumor, and ⑤ inability to obtain relevant contrast imaging and clinical data. All eligible patients underwent ^99mTc-PSMA SPECT/CT and mpMRI at an average interval of 3 days (1–7 days). None of the patients received antineoplastic therapy between the two scans. After both scans were completed, a transrectal needle prostate biopsy was performed. The patient characteristics are presented in Table 1 .

The PSMA lyophilized kit (HYNIC-PSMA) (patent number,Zl202010878750.4) was provided by the Shanghai Engineering Research Centre of Molecular Imaging Probes. The synthesis procedure has been reported previously (9). The radiochemical purity was>95%. All patients were injected intravenously with a dose of 0.74 GBq (20 mCi) ^99mTc-PSMA. Whole-body planar imaging and regional (neck-pelvic) SPECT/CT were performed 2 h after injection using a Discovery NM/CT 670Pro (GE, USA) with low-energy, high-resolution collimators. The image acquisition protocol was as follows: planar imaging: peak energy 140 keV (^99mTc) and scan velocity 15 cm/min in a 256×1025 matrix. Regional SPECT/CT: camera matrix size 128×128, zoom 1.0, rotation 360°, and 30 s/frame for 60 frames. Low-dose CT (130 keV; 60 mA) was used.

mpMRI was performed with a high-field system (Magnetom Prisma 3.0T, Siemens, The Germany) using a standardized protocol with pelvic external phased-array coils. The sequences included: transverse T1-weighted imaging (T1WI) (repetition time [TR]=500ms, echo time [TE]=12ms, field-of-view [FOV]=20 cm×20 cm, matrix=320×256); T2-weighted imaging (T2WI) (TR=5800ms, TE=106ms, FOV=20 cm×20 cm, matrix=320×256); fat-suppression spectral presaturation attenuated inversion recovery-T2WI (TR=5800ms, TE=97ms, FOV=20 cm×20 cm, matrix=320×240), and diffusion weighted imaging (DWI) (TR=5100ms, TE=64ms, FOV=20 cm×20 cm, matrix=114×114, b=50s/mm², 600 s/mm², 1500 s/mm², 3000 s/mm²). The section thickness of each sequence was 3.5 mm.

^99mTc-PSMA SPECT/CT and mpMRI images were independently read by two nuclear medicine physicians and two radiologists, respectively. The readers were blinded to the mpMRI and ^99mTc-PSMA SPECT/CT clinical reports and other readers’ findings. ^99mTc-PSMA SPECT/CT and mpMRI were performed on a workstation (Xeleris, General Electric, Waukesha, WI) and (syngo.via, Siemens Healthineers, respectively). The locations of lesions on mpMRI and ^99mTc-PSMA SPECT/CT images were compared, and lesions with the same locations on the two scans were selected as the primary lesion to extract parameters for analysis.

On mpMRI, combined with the reconstructed apparent diffusion coefficient (ADC) images, a lesion with a prostate imaging reporting and data system (PI-RADS) score > 3 was considered a positive lesion (PCa) (13). The lesions’ region of interest (ROI) was delineated, and the lowest ADC (ADC_min) was calculated. On SPECT/CT, areas with higher imaging agent uptake than normal prostate tissue after excluding physiological uptake were considered positive lesions (PCa). For imaging-based quantification analysis, Q.Metrix software (Q.Metrix GE Healthcare) was used (14). Acquisition information, including camera sensitivity, activities in full and empty syringes, administration time, and scan time, was input into the system. The volume of interest (VOI) was delineated, and the NM was 0.4. The calculated maximum standardized uptake value (SUV) voxel volume was 3.2×10^-3 mL. VOI-related quantitative parameters were automatically generated, and SUV_max was used for quantitative analysis.

On mpMRI, ①lymph node metastases: round, short-axis diameter>8 mm, uneven signals in lymph nodes on T2WI, irregular boundaries, and evident enhancement on dynamic contrast-enhanced (DCE) (15); ② bone metastases: low signal intensity on T1WI and T2WI, limited diffusion on DWI, and early enhancement after contrast agent injection on DEC (16, 17). On SPECT/CT, ①lymph node and bone metastases: uptake than normal tissue(lesion SUV_max≥liver SUV_max) after excluding lacrimal glands, salivary glands, kidneys, bladder and intestines physiological uptake. The SUV_max of all focal SPECT-positive sites was determined based on the ROI basis (17). In the quantification analysis, the size of each SPECT-positive bone and lymph node correlated with the SUV_max.

Prostate needle biopsies were performed in all participants. We used a protocol for transperineal MRI/PSMA-ultrasound fusion targeted and systematic biopsy. In brief, the image-guide (cognitive guidance, MRI/US and PSMA/US) technique was used. Targeted and systematic biopsies were performed in the same session. The number of biopsy cores was as follows: 3-4 cores for targeted biopsy and 10-12 cores for systematic biopsy. If the biopsy results were positive, patients with surgical indications underwent radical prostatectomy, and the pathological results were based on the gross specimen. For patients without surgical indications, pathological results were based on biopsy results. If the needle biopsy results are negative and the clinical symptoms are highly indicative of PCa, the patient’s serum PSA value and imaging (mpMRI, ^99mTc-PSMA SPECT/CT) should be followed up for 3-6 months. If the disease does not progress, PCa could be excluded. If the disease progresses, an additional needle biopsy should be performed (12). Not all bone and lymph node lesions showed positive pathological results. Thus, the validated method reported in previous studies was used (9, 17). All patients were followed up for at least 6 months (or until death). Serum PSA levels were reviewed every 3 months for all patients. The subsequent therapeutic schedule options depended on the patient’s condition, including radical prostatectomy, local radiation therapy, and chemotherapy. Future imaging modalities were selected according to their respective clinical needs and were not bound by a specific protocol. Patients who met at least one of the following conditions were metastases: ① Response to therapy (hormone therapy and radiation) and subsequent serum PSA decline were confirmed by follow-up examination (MRI, CT, PET, etc.); ② two or more imaging examinations recommended metastases, and ③ PSA≥100 ng/mL, suggesting distant metastases (18).

Data analysis was performed using SPSS 19.0 software (statistical product and service solutions, Chicago, Illinois). McNemar’s test was used to compare the cancer detection concordance rates between ^99mTc-PSMA SPECT/CT and mpMRI. The Mann-Whitney U test was used to compare the differences in quantitative diagnostic parameters among the different groups. Receiver operating characteristic (ROC) analysis was performed to evaluate the sensitivity, specificity, area under the ROC curve (AUC), and a cut-off value of each parameter. The Kruskal-Wallis test was used to compare the differences in quantitative diagnostic parameters among different tumor size groups. The correlation between the Gleason Score and SUV_max, ADC_min, and SUV_max/ADC_min was evaluated using Spearman correlation analysis. Logistic regression analysis was used to calculate predictors of the Gleason score. P<0.05 was considered statistically significant.

Among the 56 participants, 44 (78.5%) were diagnosed with PCa, and 12 (21.5%) with prostate hyperplasia (BPH). A flowchart illustrating the participant inclusion procedure is shown in Figure 1 . Among the 44 patients with PCa, one (2%) had neuroendocrine carcinoma, and 43 (98%) had adenocarcinoma. The surgical indications were judged by the urological surgeon according to the clinical status of the patient (2).The 23 patients with PCa diagnosed by puncture underwent robot-assisted laparoscopic radical prostatectomy (RP); postoperative pathology results were consistent with those of puncture in 11 (11/23,47.8%) patients. 12 (12/23,52.2%) patients with PCa experienced pathological upgrading. The Gleason score of patients who underwent surgery was based on the surgical specimen, and the Gleason score of patients who did not undergo surgery was based on the puncture specimen. Among 44 patients with PCa, 23 (52.3%) had metastases.

For all 56 patients, the sensitivity of ^99mTc-PSMA SPECT/CT and mpMRI in detecting primary PCa was 97.7% (43/44) and 90.9% (40/44), respectively; the difference was not statistically significant (χ^2 = 0.102, P=0.749). Their specificity was 75.0% (9/12) and 75.0% (9/12), respectively, with no statistically significant difference (χ^2 = 1.333, P=0.248) ( Table 2 ). ROC curve analysis revealed an accuracy, as measured by AUC, of 97.4% (95% CI. 93.7%-100.0%) for ^99mTc-PSMA, 95.1% (95% CI. 88.8%-100.0%) for mpMRI, and 98.2% (95% CI. 95.2%-100.0%) for ^99mTc-PSMA+mpMRI ( Figure 2 ).

The Mann-Whitney U test was used to compare differences in quantitative diagnostic parameters among the different groups ( Table 2 ). The SUV_max/ADC_min of the PCa group was significantly higher than that of the BPH group, and the SUV_max/ADC_min of the subgroup with metastases was higher than that of the subgroup without metastasis ( Table 2 ). In ^99mTc-PSMA SPECT/CT combined with mpMRI, when the cut-off value for SUV_max/ADC_min was set at 7.0×10³, the sensitivity and specificity of SUV_max/ADC_min in PCa were 93.2% (95%CI.85.7%-100.0%) and 100.0% (95%CI.100.0%-100.0%), respectively, with a Youden index of 0.932 and an AUC of 0.982 (95% CI. 95.2%-100.0%) ( Figure 3 ). When the cut-off value for SUV_max/ADC_min was set at 27.0×10³, the sensitivity and specificity of SUV_max/ADC_min in PCa with metastases was 76.2% (95%CI.58.0%-99.4%) and 73.9% (95%CI.56.0%-91.9%), respectively, with a Youden index=0.501 and AUC=0.760 (95% CI. 61.0%-91.0%) ( Figure 3A ).

The 44 prostatic lesions detected were grouped according to their maximum tumor diameter: G1 (7/44, maximum diameter < 1.0 cm), G2 (23/44, maximum diameter: 1.0 cm-3.0 cm), and G3 (14/44, maximum diameter > 3.0 cm). The Kruska–Wallis test was used to compare the differences in quantitative diagnostic parameters among different tumor size groups. There were differences in SUV_max/ADC_min among the tumor size groups; the larger the tumor size, the larger the SUV_max/ADC_min value ( Figure 4 ).

Gleason scoring is unsuitable for treating neuroendocrine PCa (19). Therefore, 43 patients with PCa were enrolled in this cohort study. Spearman correlation analysis was used, and the results revealed that ADC_min showed a weak negative correlation with Gleason score (r=-0.35, P=0.023), whereas SUV_max (r=0.61, P=0.008) and SUV_max/ADC_min (r=0.59, P=0.023) showed a moderate positive correlation with Gleason score ( Figure 5 ). Based on the Gleason score, the patients were divided into high-(Gleason score>7, n=31), medium- (Gleason score=7a, n=4;Gleason score=7b, n=2) and low-risk groups (Gleason score < 7, n=6). According to the Mann-Whitney U test, there were statistical differences in SUV_max, ADC_min, and SUV_max/ADC_min between the high-, medium- and low-risk groups (all P < 0.05) ( Table 3 ). With the presence of a high-risk group (yes=1, no=0) as the dependent variable, and age, serum PSA level, and SUV_max/ADC_min as the independent variables, logistic regression analysis showed that SUV_max/ADC_min was independently correlated with the presence of a high-risk group; for every 1×10³ increase in SUV_max/ADC_min, the detection rate of the high-risk group increased by 55.7% (OR=1.557, P=0.042). When the cut-off value for SUV_max/ADC_min was set at 15.0×10³, the sensitivity and specificity of SUV_max/ADC_min in the high-risk PCa group were 78.4% (95%CI.65.1%-91.6%) and 100.0% (95%CI.100.0%-100.0%), respectively, with a Youden index of 0.784 and AUC of 0.928 (95% CI. 84.0%-100.0%) ( Figure 3B ) ( Table 4 ).