RPS 916 - The evolving landscape of hybrid imaging for oncology

Author Block: T. Spiegel¹, D. A. Resch¹, S. Rasul¹, O. Lafcı¹, N. Pötsch¹, P. Clauser¹, P. A. Baltzer¹, K. Pinker-Domenig², T. H. Helbich¹; ¹Vienna/AT, ²New York, NY/US
Purpose: 18F-Fluoroestradiol (18F-FES) is a novel radiotracer primarily used for staging ER-positive breast cancer (BC). However, understanding its uptake characteristics in both benign and malignant breast lesions is essential. This study evaluates 18F-FES-PET/MRI uptake patterns across receptor status, molecular BC subtypes, and benign breast lesions.
Methods or Background: This retrospective analysis of prospectively acquired data includes 41 women with 48 breast lesions (42 malignant, 6 benign) who underwent 18F-FES-PET/MRI. All patients underwent simultaneous 18F-FES-PET/MRI performed using a Biograph mMR system (Siemens, Germany). Quantitative PET parameters, including SUVmax, and lesion size were assessed. Histopathology in malignant tumors and histopathology or follow-up >2 years in benign lesions was the reference standard. Mann-Whitney U test and ROC analyses were performed; p<0.05 was considered significant.
Results or Findings: Among 42 BCs, 39 were ER-positive (8 Luminal A, 31 Luminal B) and 3 triple-negative breast cancers (TNBC). ER-positive BCs had statistically significantly higher SUVmax (median 2.40, IQR 1.31-3.59) than ER-negatives (SUVmax <1.00, p=0.014). 6 ER-positive BCs with lesion sizes <9 mm failed to show any uptake. AUC for SUVmax in lesions ≥10 mm to differentiate between ER-positive and ER-negative BCs was 1.00 and when including all lesions AUC 0.902. SUVmax tended to be higher in Luminal B than Luminal A (p=0.173), both exceeding TNBC (p<0.001 and p=0.036). Benign lesions (n=6) showed SUVmax values between 0.72-1.57, statistically significantly lower than in BCs (p<0.001).
Conclusion: 18F-FES uptake is predominantly observed in ER-positive breast cancers and reliably distinguishes ER-positive from ER-negative tumors ≥10 mm. However, overlap exists among molecular subtypes, particularly Luminal A and B, and benign lesions may also show variable uptake.
Limitations: Retrospective design and small sample sizes in certain subgroups.
Funding for this study: OeNB Anniversary Fund AB18207ONB
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: EK Nr. 510/2009

Author Block: E. Davydova, M. A. Komarova, V. Е. Sinitsyn, A. Martinovich, O. Ryzhova; Moscow/RU
Purpose: To compare the diagnostic accuracy of FAPI PET/CT and FDG PET/CT in patients with liver tumor.
Methods or Background: 15 patients (mean age 57,5 y.o, 7male/8 female) diagnosed with liver cancer who underwent FAPI and FDG PET/CT between 7 days were included in the study. 4pts were examined to detect progression of the underlying disease (had the previous treatment). The all liver lesions had morphological verification (11pts had cholangiocarcinoma G1-3, 2pts. - hepatocellular carcinoma, 1- intraductal papillary tumor of the bile duct and 1pts - calcifying nested stromal-epithelial tumor). Uptake intensity(SUVmax) and target-to-background ratio(TBR) of the primary liver tumor and lymph nodes metastasis were compared between two imaging modalities. Sensitivity(Se), specificity(Sp), positive predictive value(PPV), negative predictive value(NPV) and accuracy(Ac) were calculated for each modality on a per-region basis.
Results or Findings: 14 pts were found primary tumor and 3 pts of them had metastases in regional lymph nodes or lung. There was a better contrast with 68Ga-FAPI, with SUVmax of the liver tumors of 14,9 [8,94-23,48] compared with SUVmax of 7,81 [3,55-10,21] of liver tumors with 18F-FDG (p=0,02). The TBR of primary tumor were higher in 68Ga-FAPI compared with 18F-FDG - 10,56[4,15-16,46] vs. 2,5[1,17-3,67] respectively, p=0,0014. For detection of primary liver tumor and metastasis on region based level Se, Sp, PPV, NPV and Ac of Ga-68 FAPI PET/CT were higher than F18-FDG PET/CT (97,3%, 71,43 %, 90,00%, 90,91% and 90,20% vs. 78,38%, 57,14%, 82,86%,50,00% and 72,55 %, respectively).
Conclusion: FAPI PET-CT is a very promising tool for evaluation of patients with liver tumor because more accurate than FDG PET-CT. High level of FAPI uptake in the primary hepatic lesions is indicate the potential of FAP-targeted radionuclide therapy.
Limitations: No
Funding for this study: Without funding
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: N. A. Umretiya, V. Mahajan, R. Arafath; Chennai/IN
Purpose: To evaluate the diagnostic performance of the proposed Axillary Node Reporting and Data System (AN-RADS) ultrasound classification for axillary nodes in breast cancer staging and to correlate AN-RADS categories with PET-CT and histopathology.
Methods or Background: This prospective observational study enrolled 15 patients with biopsy-proven breast carcinoma undergoing baseline staging between 1st August and 1st September 2025. All patients underwent axillary ultrasound with AN-RADS scoring (categories 1–5) and 18F-FDG PET-CT. PET positivity was defined as SUVmax ≥2.5. Histopathology (SLNB, core, or excision biopsy) or FNAC was available in 13/15 patients; 2 were excluded from diagnostic accuracy calculations due to lack of biopsy. AN-RADS ≥4 was considered ultrasound-positive. Statistical analysis included malignancy rates per category, diagnostic accuracy metrics with exact 95% confidence intervals, and Cohen’s κ for concordance.
Results or Findings: Median patient age was 54 years (range: 42–68). AN-RADS distribution was: category 2 in 5 patients (33.3%), category 3 in 1 (6.7%), category 4 in 3 (20.0%), and category 5 in 6 (40.0%). Among 13 evaluable cases, 8/13 (61.5%; 95% CI 31.6–86.1) demonstrated metastatic nodes. Malignancy rates were 0% in AN-RADS 2 and 100% in AN-RADS 4–5. With AN-RADS ≥4 as threshold, ultrasound achieved sensitivity, specificity, PPV, NPV, and accuracy of 100%. PET-CT using SUVmax ≥2.5 showed identical diagnostic performance. Ultrasound, PET-CT, and histopathology showed perfect concordance (κ = 1.0).
Conclusion: The proposed AN-RADS system demonstrates a clear malignancy gradient and excellent correlation with PET-CT and histopathology, suggesting its potential as a structured, reliable framework for axillary nodal staging in breast cancer.
Limitations: Small sample size, short study period, and selection bias limit generalizability. The perfect performance observed is likely influenced by the restricted dataset.
Funding for this study: No external funding was received.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: R. Regazzo, A. Cattabriga, C. Gaudiano, A. Bardelli, L. Bianchi, A. Farolfi, S. Fanti, R. Schiavina, C. Mosconi; Bologna/IT
Purpose: PSMA-PET is the gold standard for staging and restaging high-risk prostate cancer (PCa), while whole-body MRI (wb-MRI) may represent a radiation-free alternative despite limited evidence. This prospective single-center study aimed to assess the diagnostic non-inferiority of wb-MRI compared to PSMA-PET for metastatic evaluation in high-risk patients at primary treatment and in those with biochemical recurrence (BCR).
Methods or Background: We enrolled 89 PCa patients: 47 high/very high-risk men undergoing radical prostatectomy with extended pelvic lymph node dissection (RARP + ePLND, Group 1) and 42 men with PSA persistence or BCR after surgery (Group 2). All patients underwent PSMA-PET and wb-MRI within 30 days. Reference standards were histopathology in Group 1 and clinical/imaging follow-up in Group 2. Sensitivity, specificity, PPV, NPV, accuracy, and AUC from ROC analyses were compared between modalities.
Results or Findings: In Group 1, pathological N1 disease was found in 42%. PSMA-PET vs wb-MRI for nodal staging showed sensitivity 53% vs 41%, specificity 90% vs 87%, and AUC 0.71 vs 0.64, with no significant differences.
In Group 2, 21% had PSA persistence and 79% BCR (median time 28 months, median PSA 0.3 ng/mL). PSMA-PET vs wb-MRI detected local recurrence in 17% vs 29%, nodal disease in 12% vs 19%, and distant metastases in 33% vs 19%. Diagnostic performance was comparable (sensitivity 54% vs 46%, specificity 86% vs 89%, AUC 0.80 vs 0.78).
Conclusion: Wb-MRI demonstrates diagnostic performance comparable to PSMA-PET for staging and restaging high-risk PCa and may represent a valid alternative in selected patients.
Limitations: Preliminary single-center results; limited sample size; heterogeneous follow-up in the BCR group.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: F. Abubacker Sulaiman¹, A. Asokan², S. Sundaraiya¹; ¹Chennai/IN, ²Coimbatore/IN
Purpose: To assess the prognostic and predictive value of baseline metabolic tumor volume (MTV) and total lesion glycolysis (TLG) derived from ^18F-FDG PET/CT in patients with solid malignancies, and to correlate these volumetric biomarkers with treatment response and progression-free survival (PFS).
Methods or Background: A prospective study was performed on 75 patients with biopsy-proven malignancies, including gynecologic, gastrointestinal, and thoracic cancers, who underwent baseline ^18F-FDG PET/CT prior to therapy. Quantitative parameters—SUVmax, SUVmean, MTV, and TLG—were obtained using semi-automated segmentation (SUV threshold ≥41% of SUVmax). Treatment response was assessed using PERCIST criteria at 3 and 6 months. Patients were followed up for 12 months to evaluate PFS. Statistical analysis included Kaplan–Meier survival curves, Cox regression, and Pearson’s correlation.
Results or Findings: Baseline MTV and TLG values demonstrated strong predictive correlation with treatment outcomes and PFS. Patients with high MTV (>120 cm³) and TLG (>850 g·cm³) showed significantly lower response rates and shorter PFS (median 6.4 months) compared to those with low MTV/TLG (median 14.8 months; p < 0.001). While SUVmax alone failed to predict long-term outcome, combined volumetric parameters (MTV + TLG) achieved the highest prognostic accuracy (AUC = 0.91). Multivariate analysis confirmed MTV and TLG as independent predictors of disease progression.
Conclusion: Baseline metabolic tumor volume and total lesion glycolysis are robust quantitative biomarkers for predicting treatment response and survival, outperforming conventional SUV metrics in oncologic PET/CT.
Limitations: Single-center design, heterogeneous tumor population, and short follow-up duration limit generalizability. Multicentric studies with standardized segmentation protocols are warranted.
Funding for this study: No external funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Institutional ethical committee approval was obtained.

Author Block: N. C. Balci, G. Iuppa, A. K. Ansari, C. Hajj, A. Khan; Abu Dhabi/AE
Purpose: FDG-PETCT derived MTV is determinant parameter in oncologic assessment in colon cancer liver metastases. This study aims to identify if DWI can provide correlative results.
Methods or Background: FDG PETCT and DWI MRI of 50 patients with colon cancer liver metastases were retrospectively enrolled in the study, PETCT derived MTV was measured using automated software . DWI MRI was acquired in the scanner with highest b value of 800 mm2/sec and synthetic higher b value images were created in workstation with the values of 1000 mm2/sec, 1500 mm2/sec and 2000 mm2/sec. For each patient, the total number of liver lesions, total MTV and total lesion volumes on DWI MRI were calculated both on PETCT and on four different b value images. Correlation of total liver lesions was made with Pearson’s r test. Correlation of MTV with DWI was made using paired t-test.
Results or Findings: Average number lesions for each patient were: 3.04 for PETCT, 3.67 for b= 800 mm2/sec (Pearson r=0.97), 3.45 for b=1000 mm2/sec (Pearson r=0.98), 3.07 for b=1500 mm2/sec (Pearson r= 0.99) and 2.97 for b=2000 mm2/sec (p Pearson r=0.98). The average total MTV was 56.25 cc, average total tumor volumes on DWI were 59.98 cc for b=800 mm2/sec, p < 0.001 (significantly higher than PET-CT); 58.41 cc for b=1000 mm2/sec, p < 0.001 (significantly higher than PET-CT); 56.41 cc for b=1500 mm2/sec, p = 0.021 (least significant difference); and 54.22cc for b=2000 mm2/sec, p < 0.001 (significantly lower than PET-CT).
Conclusion: The result of this study reveals significant correlation of number of lesions and total tumor volume compared to PETCT lesion count and MTV when b value of 1500 is assessed on DWI.
Limitations: Histopathology correlation of all metastases missing.
Funding for this study: Imaging Institute Cleveland Clinic
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Research Ethics Committee Cleveland Clinic Abu Dhabi

Author Block: J. Bai, J. Lu; Beijing/CN
Purpose: To evaluate the diagnostic potential of multiparametric 18F-FET PET/MRI (mpPET/MRI) features in distinguishing tumor recurrence from radiation necrosis (RN) in brain metastases (BMs) following Gamma Knife radiosurgery (GKRS) using machine learning techniques.
Methods or Background: PET/MRI data were obtained from 74 patients with 114 suspected lesions after GKRS included CE-T1WI, ADC, CBF, and FET PET. We extracted 1,080 original and transformed first-order features from these specified modalities using the eXtreme Gradient Boosting (XGBoost) classifier and Spearman correlation analysis. Shapley Additive Explanations (SHAP) evaluated their significance. Six machine learning classifiers were built based on mpPET/MRI data. The diagnostic performance of the mpPET/MRI model and each single-modality submodel was compared using the area under the curve (AUC).
Results or Findings: Fifteen important first-order features (6 from FET PET, 3 from CE-T1WI, 3 from ADC, and 3 from CBF) were selected. XGBoost was chosen as the optimal classifier. The diagnostic performance of the mpPET/MRI model (training/validation AUC = 0.93/0.92) was higher than that of CE-T1WI (AUC = 0.85/0.82), ADC (AUC = 0.86/0.84), CBF (AUC = 0.86/0.83) and FET PET (AUC = 0.87/0.87) models. Moreover, the mpPET/MRI model improves diagnostic accuracy by 14% over the standard Response Assessment in Neuro-Oncology criteria (89.90% vs. 75.43%). SHAP analysis highlighted significant predictors such as log- and wavelet-transformed first-order features of FET PET.
Conclusion: The XGBoost model developed from mpPET/MRI data may be a promising approach to identify the distribution of RN and recurrent tumor in patients with BMs after GKRS. Particularly, FET PET played a significant role.
Limitations: The limited sample size and single-center design of this study restrict subgroup analyses and generalizability. Future work should incorporate advanced machine learning techniques such as texture analysis and ensemble learning to enhance model predictive performance.
Funding for this study: Huizhi Ascent Project of Xuanwu Hospital [HZ2021ZCLJ005]
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Approved by the Ethics Committee and Institutional Review Board of Xuanwu Hospital, Capital Medical University (No. 2023-044).

Author Block: A. Cattabriga, M. Talarico, C. Nanni, R. Regazzo, G. Pertile, S. Brocchi, S. Fanti, E. Zamagni, C. Mosconi; Bologna/IT
Purpose: The aim of this study was to compare the diagnostic performance of PET/CT and WBMRI in the diagnosis and staging of smoldering multiple myeloma (SMM) and newly diagnosed multiple myeloma (NDMM), to correlate both imaging modalities with markers of disease burden and prognostic features and to validate in clinical practice the imaging response criteria IMPeTUs and MY-RADS.
Methods or Background: 205 patients (74 SMM, 131 NDMM) were prospectively enrolled between October 2022 and April 2025. All underwent baseline assessment with WBMRI and FDG-PET/CT. In NDMM, baseline-positive imaging was repeated before maintenance in transplant-eligible patients and after one year of treatment in transplant-ineligible patients. Deauville/IMPeTUs and MY-RADS criteria were applied to define focal lesions (FLs), diffuse disease and treatment response. Imaging findings were correlated with biochemical parameters and prognostic scores (ISS/R-ISS).
Results or Findings: FLs and paraskeletal disease were detected more frequently by WBMRI vs PET/CT (p<0.05). Both modalities showed concordance in detecting extramedullary disease (κ=1). WBMRI-detected FLs significantly correlated with worse prognosis by ISS/R-ISS. Diffuse disease was detected more frequently by PET/CT (p<0.05). However, WBMRI-defined diffuse disease correlated with multiple clinical parameters (hemoglobin <10 g/dL, bone marrow plasma cells, paraprotein, urinary protein), whereas PET-defined diffuse disease did not. In SMM, detection of FLs and diffuse disease was similar across techniques. In the preliminary analysis of the first 45 reassessed patients, WBMRI and PET/CT showed good concordance in response assessment (κ=0.69).
Conclusion: WBMRI was superior to PET/CT in detecting FLs and paraskeletal disease and correlated better with disease burden and prognosis. Both modalities showed concordance for extramedullary disease and response assessment.
Limitations: Main limitations are the small sample size in some subgroups (extramedullary disease, post-treatment reassessment). These cohorts are expected to expand; updated data will be presented.
Funding for this study: This study received no external funding. All expenses were covered by our institution’s research funding
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The institutional ethics board of our institution (IRCCS Policlinico di Sant’Orsola, Bologna) approved this study prior to patient enrollment

Author Block: M. Garcia Fontes, M. RODRIGUEZ PARODI, L. Valuntas, E. Otero, L. Mouro, V. Gigirey, L. Servente, G. Dos Santos; Montevideo/UY
Purpose: The detection of suspicious lesions of prostate cancer using Multiparametric Magnetic Resonance(mpMRI) is challenging because of the many pitfalls that often lead to unnecessary biopsies.
The aim of the present study is to compare the concordance of the PIRADS Scale V 2.1 of mpMRI with the PRIMARY SCORE of the PSMA PET CT in the detection of suspicious lesions of clinically significative prostate cancer to determine if it can add sensitivity or specificity in order to guide more effective biopsies.
Methods or Background: Fifty patients aged between 48 and 85 years were included. They underwent mpMRI using 3T equipment, fusion biopsy using MRI and ultrasound, and PSMA PET/CT for initial staging.
The number of PIRADS lesions detected by mpMRI were compared with the uptake areas using the PRIMARY PSMA Score and both methods were correlated with the pathological anatomy findings.
Results or Findings: A statistically significant correlation was found between PIRADS 4 and 5 lesions on mpMRI and SCORE 3 and 4 lesions on PSMA PET for Gleason 7 and higher grades found in the pahtologic anatomy.
In one patient with a hip prosthesis, the PET PSMA Primary Score allowed the detection of suspicious lesions that were very doubtful due to artifacts in the mpMRI.
Conclusion: The analysis of multiparametric prostate MRI using the PIRADS V2.1 and the PSMA PET Primary Score has similar sensitivity for detecting suspicious lesions for prostate cancer to guide prostate biopsy.
The PET PSMA Primary Score has slightly higher specificity than PIRADS V2.1 of mpMRI for clinically significant cancers.
Both methods could be use together to detect suspicious lesions of prostate cancer increasing sensitivity.
In patients with contraindications for mpMRI, PSMA PET SCORE seems to be a useful tool.
Limitations: Not aplicable
Funding for this study: No funding was provided for this study
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Yes