European Journal of Nuclear Medicine and Molecular Imaging最新文献

Background

Pathological grade is a critical determinant of clinical outcomes and decision-making of follicular lymphoma (FL). This study aimed to develop a deep learning model as a digital biopsy for the non-invasive identification of FL grade.

Methods

This study retrospectively included 513 FL patients from five independent hospital centers, randomly divided into training, internal validation, and external validation cohorts. A multimodal fusion Transformer model was developed integrating 3D PET tumor images with tabular data to predict FL grade. Additionally, the model is equipped with explainable modules, including Gradient-weighted Class Activation Mapping (Grad-CAM) for PET images, SHapley Additive exPlanations analysis for tabular data, and the calculation of predictive contribution ratios for both modalities, to enhance clinical interpretability and reliability. The predictive performance was evaluated using the area under the receiver operating characteristic curve (AUC) and accuracy, and its prognostic value was also assessed.

Results

The Transformer model demonstrated high accuracy in grading FL, with AUCs of 0.964–0.985 and accuracies of 90.2-96.7% in the training cohort, and similar performance in the validation cohorts (AUCs: 0.936–0.971, accuracies: 86.4-97.0%). Ablation studies confirmed that the fusion model outperformed single-modality models (AUCs: 0.974 − 0.956, accuracies: 89.8%-85.8%). Interpretability analysis revealed that PET images contributed 81-89% of the predictive value. Grad-CAM highlighted the tumor and peri-tumor regions. The model also effectively stratified patients by survival risk (P < 0.05), highlighting its prognostic value.

Conclusions

Our study developed an explainable multimodal fusion Transformer model for accurate grading and prognosis of FL, with the potential to aid clinical decision-making.

Graphical Abstract

Purpose

To evaluate the diagnostic accuracy and clinical impact of fibroblast activation protein (FAP)-targeted PET/CT imaging in primary and metastatic breast cancer and compare the results with those of standard-of-care imaging (SCI) and [¹⁸F]FDG PET/CT.

Methods

We prospectively analyzed patients with diagnosed or suspected breast cancer who underwent concomitant FAP-targeted PET/CT (radiotracers including either [⁶⁸Ga]Ga-FAPI-46 or [¹⁸F]FAPI-42) and [¹⁸F]FDG PET/CT scans from June 2020 to January 2024 at two medical centers. Breast ultrasound (US) imaging was performed in all treatment-naïve patients as SCI. The SUVmax, tumor-to-background ratio (TBR), lesion detection rate, and tumor-node-metastasis (TNM) classifications between FAP-targeted and [¹⁸F]FDG PET/CT were evaluated and compared.

Results

Sixty-one female patients (median age, 52 y; range, 28–82 y) were included. Among them, 23 patients underwent evaluation for a definitive diagnosis of suspected breast lesions, 15 underwent initial staging, and 23 were evaluated for the detection of recurrence. The sensitivities of breast US, [¹⁸F]FDG, and FAP-targeted PET/CT for detecting primary breast tumors were 82%, 79%, and 100%, respectively. Regarding the diagnosis of recurrent/metastatic lesions, the lesion-based detection rate of FAP-targeted PET/CT was significantly higher than that of [¹⁸F]FDG, which included local and regional recurrence, neck lymph node (LN), abdomen LN, bone, and liver metastases. Compared with [¹⁸F]FDG PET/CT, FAP-targeted PET/CT altered thirteen patients’ TNM staging/restaging (13/59, 22%) and nine patients’ clinical management (9/59, 15%). Compared to SCI, FAPI changed fourteen patients’ TNM staging/re-staging (14/59, 24%) and eleven patients’ therapeutic regimens(11/59, 19%). There was no significant association between FAPI-derived SUVmax and receptor status/histologic type in both primary and metastatic lesions.

Conclusion

FAP-targeted PET/CT was superior to [¹⁸F]FDG in diagnosing primary and metastatic breast cancer, with higher radiotracer uptake and TBR, especially in the detection of primary/recurrent tumors, abdominal LN metastases, liver, and bone metastases. FAP-targeted PET/CT is superior to [¹⁸F]FDG and SCI in TNM staging and may improve tumor staging, recurrence detection, and implementation of necessary treatment modifications.

Purpose

The positron range effect can impair PET image quality of Gallium-68 (⁶⁸Ga). A positron range correction (PRC) can be applied to reduce this effect. In this study, the effect of a tissue-independent PRC for ⁶⁸Ga was investigated on patient data.

Methods

PET/CT data (40 patients: [⁶⁸Ga]Ga-DOTATOC or [⁶⁸Ga]Ga-PSMA) were reconstructed using Q.Clear reconstruction algorithm. Two reconstructions were performed per patient, Q.Clear with and without PRC. SUV_max and contrast-to-noise ratio (CNR) values per lesion were compared between PRC and non-PRC images. Five experienced nuclear medicine physicians reviewed the images and chose the preferred reconstruction based on the image quality, lesion detectability, and diagnostic confidence.

Results

A total of 155 lesions were identified. The PRC resulted in statistically significant increase of the SUV_max and CNR for soft tissue lesions (6.4%, p < 0.001; 8.6%, p < 0.001), bone lesions (14.6%, p < 0.001; 12.5%, p < 0.001), and lung lesions (3.6%, p = 0.010; 6.3%, p = 0.001). This effect was most prominent in small lesions (SUV_max: 12.0%, p < 0.001, and CNR: 13.0%, p < 0.001). Similar or better image quality, lesion detectability, and diagnostic confidence was achieved in PRC images compared to the non-PRC images as those assessed by the expert readers.

Conclusions

A tissue-independent PRC increased the SUV_max and CNR in soft tissue, bone, and lung lesions with a larger effect for the small lesions. Visual assessment demonstrated similar or better image quality, lesion detectability, and diagnostic confidence in PRC images compared to the non-PRC images.

Purpose

Advancements of deep learning in medical imaging are often constrained by the limited availability of large, annotated datasets, resulting in underperforming models when deployed under real-world conditions. This study investigated a generative artificial intelligence (AI) approach to create synthetic medical images taking the example of bone scintigraphy scans, to increase the data diversity of small-scale datasets for more effective model training and improved generalization.

Methods

We trained a generative model on ^99mTc-bone scintigraphy scans from 9,170 patients in one center to generate high-quality and fully anonymized annotated scans of patients representing two distinct disease patterns: abnormal uptake indicative of (i) bone metastases and (ii) cardiac uptake indicative of cardiac amyloidosis. A blinded reader study was performed to assess the clinical validity and quality of the generated data. We investigated the added value of the generated data by augmenting an independent small single-center dataset with synthetic data and by training a deep learning model to detect abnormal uptake in a downstream classification task. We tested this model on 7,472 scans from 6,448 patients across four external sites in a cross-tracer and cross-scanner setting and associated the resulting model predictions with clinical outcomes.

Results

The clinical value and high quality of the synthetic imaging data were confirmed by four readers, who were unable to distinguish synthetic scans from real scans (average accuracy: 0.48% [95% CI 0.46–0.51]), disagreeing in 239 (60%) of 400 cases (Fleiss’ kappa: 0.18). Adding synthetic data to the training set improved model performance by a mean (± SD) of 33(± 10)% AUC (p < 0.0001) for detecting abnormal uptake indicative of bone metastases and by 5(± 4)% AUC (p < 0.0001) for detecting uptake indicative of cardiac amyloidosis across both internal and external testing cohorts, compared to models without synthetic training data. Patients with predicted abnormal uptake had adverse clinical outcomes (log-rank: p < 0.0001).

Conclusions

Generative AI enables the targeted generation of bone scintigraphy images representing different clinical conditions. Our findings point to the potential of synthetic data to overcome challenges in data sharing and in developing reliable and prognostic deep learning models in data-limited environments.

Graphical abstract

Purpose

The study explores the role of multimodal imaging techniques, such as [¹⁸F]F-PSMA-1007 PET/CT and multiparametric MRI (mpMRI), in predicting the ISUP (International Society of Urological Pathology) grading of prostate cancer. The goal is to enhance diagnostic accuracy and improve clinical decision-making by integrating these advanced imaging modalities with clinical variables. In particular, the study investigates the application of few-shot learning to address the challenge of limited data in prostate cancer imaging, which is often a common issue in medical research.

Methods

This study conducted a retrospective analysis of 341 prostate cancer patients enrolled between 2019 and 2023, with data collected from five imaging modalities: [¹⁸F]F-PSMA-1007 PET, CT, Diffusion Weighted Imaging (DWI), T2 Weighted Imaging (T2WI), and Apparent Diffusion Coefficient (ADC). The study compared the performance of five single-modality data sets, PET/CT dual-modality fusion data, mpMRI tri-modality fusion data, and five-modality fusion data within deep learning networks, analyzing how different modalities impact the accuracy of ISUP grading prediction. To address the issue of limited data, a few-shot deep learning network was employed, enabling training and cross-validation with only a small set of labeled samples. Additionally, the results were compared with those from preoperative biopsies and clinical prediction models to further assess the reliability of the experimental findings.

Results

The experimental results demonstrate that the multimodal model (combining [¹⁸F]F-PSMA-1007 PET/CT and multiparametric MRI) significantly outperforms other models in predicting ISUP grading of prostate cancer. Meanwhile, both the PET/CT dual-modality and mpMRI tri-modality models outperform the single-modality model, with comparable performance between the two multimodal models. Furthermore, the experimental data confirm that the few-shot learning network introduced in this study provides reliable predictions, even with limited data.

Conclusion

This study highlights the potential of applying multimodal imaging techniques (such as PET/CT and mpMRI) in predicting ISUP grading of prostate cancer. The findings suggest that this integrated approach can enhance the accuracy of prostate cancer diagnosis and contribute to more personalized treatment planning. Furthermore, incorporating few-shot learning into the model development process allows for more robust predictions despite limited data, making this approach highly valuable in clinical settings with sparse data.

Purpose

We hypothesised that applying radiomics to [¹⁸F]PSMA-1007 PET/CT images could help distinguish Unspecific Bone Uptakes (UBUs) from bone metastases in prostate cancer (PCa) patients. We compared the performance of radiomic features to human visual interpretation.

Materials and methods

We retrospectively analysed 102 hormone-sensitive PCa patients who underwent [¹⁸F]PSMA-1007 PET/CT and exhibited at least one focal bone uptake with known clinical follow-up (reference standard). Using matRadiomics, we extracted features from PET and CT images of each bone uptake and identified the best predictor model for bone metastases using a machine-learning approach to generate a radiomic score. Blinded PET readers with low (n = 2) and high (n = 2) experience rated each bone uptake as either UBU or bone metastasis. The same readers performed a second read three months later, with access to the radiomic score.

Results

Of the 178 [¹⁸F]PSMA-1007 bone uptakes, 74 (41.5%) were classified as PCa metastases by the reference standard. A radiomic model combining PET and CT features achieved an accuracy of 84.69%, though it did not surpass expert PET readers in either round. Less-experienced readers had significantly lower diagnostic accuracy at baseline (p < 0.05) but improved with the addition of radiomic scores (p < 0.05 compared to the first round).

Conclusion

Radiomics might help to differentiate bone metastases from UBUs. While it did not exceed expert visual assessments, radiomics has the potential to enhance the diagnostic accuracy of less-experienced readers in evaluating [¹⁸F]PSMA-1007 PET/CT bone uptakes.

Purpose

To develop and validate a prostate-specific membrane antigen (PSMA) PET/CT based multimodal deep learning model for predicting pathological lymph node invasion (LNI) in prostate cancer (PCa) patients identified as candidates for extended pelvic lymph node dissection (ePLND) by preoperative nomograms.

Methods

[⁶⁸Ga]Ga-PSMA-617 PET/CT scan of 116 eligible PCa patients (82 in the training cohort and 34 in the test cohort) who underwent radical prostatectomy with ePLND were analyzed in our study. The Med3D deep learning network was utilized to extract discriminative features from the entire prostate volume of interest on the PET/CT images. Subsequently, a multimodal model i.e., Multi kernel Support Vector Machine was constructed to combine the PET/CT deep learning features, quantitative PET and clinical parameters. The performance of the multimodal models was assessed using final histopathology as the reference standard, with evaluation metrics including area under the receiver operating characteristic curve (AUC), calibration curve, decision curve analysis, and compared with available nomograms and PET/CT visual evaluation result.

Results

Our multimodal model incorporated clinical information, maximum standardized uptake value (SUV_max), and PET/CT deep learning features. The AUC for predicting LNI was 0.89 (95% confidence interval [CI] 0.81–0.97) for the final model. The proposed model demonstrated superior predictive accuracy in the test cohort compared to PET/CT visual evaluation result, the Memorial Sloan Kettering Cancer Center (MSKCC) and the Briganti-2017 nomograms (AUC 0.85 [95% CI 0.69-1.00] vs. 0.80 [95% CI 0.64–0.95] vs. 0.79 [95% CI 0.61–0.97] and 0.69 [95% CI 0.50–0.88], respectively). The proposed model showed similar calibration and higher net benefit as compared to the traditional nomograms.

Conclusion

Our multimodal deep learning model, which incorporates preoperative PSMA PET/CT imaging, shows enhanced predictive capabilities for LNI in clinically localized PCa compared to PSMA PET/CT visual evaluation result and existing nomograms like the MSKCC and Briganti-2017 nomograms. This model has the potential to reduce unnecessary ePLND procedures while minimizing the risk of missing cases of LNI.