RESUMEN
PURPOSE: Total metabolic tumor volume (TMTV) segmentation has significant value enabling quantitative imaging biomarkers for lymphoma management. In this work, we tackle the challenging task of automated tumor delineation in lymphoma from PET/CT scans using a cascaded approach. METHODS: Our study included 1418 2-[18F]FDG PET/CT scans from four different centers. The dataset was divided into 900 scans for development/validation/testing phases and 518 for multi-center external testing. The former consisted of 450 lymphoma, lung cancer, and melanoma scans, along with 450 negative scans, while the latter consisted of lymphoma patients from different centers with diffuse large B cell, primary mediastinal large B cell, and classic Hodgkin lymphoma cases. Our approach involves resampling PET/CT images into different voxel sizes in the first step, followed by training multi-resolution 3D U-Nets on each resampled dataset using a fivefold cross-validation scheme. The models trained on different data splits were ensemble. After applying soft voting to the predicted masks, in the second step, we input the probability-averaged predictions, along with the input imaging data, into another 3D U-Net. Models were trained with semi-supervised loss. We additionally considered the effectiveness of using test time augmentation (TTA) to improve the segmentation performance after training. In addition to quantitative analysis including Dice score (DSC) and TMTV comparisons, the qualitative evaluation was also conducted by nuclear medicine physicians. RESULTS: Our cascaded soft-voting guided approach resulted in performance with an average DSC of 0.68 ± 0.12 for the internal test data from developmental dataset, and an average DSC of 0.66 ± 0.18 on the multi-site external data (n = 518), significantly outperforming (p < 0.001) state-of-the-art (SOTA) approaches including nnU-Net and SWIN UNETR. While TTA yielded enhanced performance gains for some of the comparator methods, its impact on our cascaded approach was found to be negligible (DSC: 0.66 ± 0.16). Our approach reliably quantified TMTV, with a correlation of 0.89 with the ground truth (p < 0.001). Furthermore, in terms of visual assessment, concordance between quantitative evaluations and clinician feedback was observed in the majority of cases. The average relative error (ARE) and the absolute error (AE) in TMTV prediction on external multi-centric dataset were ARE = 0.43 ± 0.54 and AE = 157.32 ± 378.12 (mL) for all the external test data (n = 518), and ARE = 0.30 ± 0.22 and AE = 82.05 ± 99.78 (mL) when the 10% outliers (n = 53) were excluded. CONCLUSION: TMTV-Net demonstrates strong performance and generalizability in TMTV segmentation across multi-site external datasets, encompassing various lymphoma subtypes. A negligible reduction of 2% in overall performance during testing on external data highlights robust model generalizability across different centers and cancer types, likely attributable to its training with resampled inputs. Our model is publicly available, allowing easy multi-site evaluation and generalizability analysis on datasets from different institutions.
Asunto(s)
Procesamiento de Imagen Asistido por Computador , Linfoma , Tomografía Computarizada por Tomografía de Emisión de Positrones , Carga Tumoral , Humanos , Tomografía Computarizada por Tomografía de Emisión de Positrones/métodos , Linfoma/diagnóstico por imagen , Procesamiento de Imagen Asistido por Computador/métodos , Fluorodesoxiglucosa F18 , Automatización , Masculino , FemeninoRESUMEN
The gastrin-releasing peptide receptor (GRPR) is overexpressed in prostate cancer and other solid malignancies. Following up on our work on [68 Ga]Ga-ProBOMB1 that had better imaging characteristics than [68 Ga]Ga-NeoBOMB1, we investigated the effects of substituting 68 Ga for 177 Lu to determine if the resulting radiopharmaceuticals could be used with a therapeutic aim. We radiolabeled the bombesin antagonist ProBOMB1 (DOTA-pABzA-DIG-D-Phe-Gln-Trp-Ala-Val-Gly-His-Leu-ψ-Pro-NH2 ) with lutetium-177 and compared it with [177 Lu]Lu-NeoBOMB1 (obtained in 54.2 ± 16.5% isolated radiochemical yield with >96% radiochemical purity and 440.8 ± 165.1 GBq/µmol molar activity) for GRPR targeting. Lu-NeoBOMB1 had better binding affinity for GRPR than Lu-ProBOMB1 (Ki values: 2.26 ± 0.24 and 30.2 ± 3.23nM). [177 Lu]Lu-ProBOMB1 was obtained in 53.7 ± 5.4% decay-corrected radiochemical yield with 444.2 ± 193.2 GBq/µmol molar activity and >95% radiochemical purity. In PC-3 prostate cancer xenograft mice, tumor uptake of [177 Lu]Lu-ProBOMB1 was 3.38 ± 1.00, 1.32 ± 0.24, and 0.31 ± 0.04%ID/g at 1, 4, and 24 hours pi. However, the uptake in tumor was lower than [177 Lu]Lu-NeoBOMB1 at all time points. [177 Lu]Lu-ProBOMB1 was inferior to [177 Lu]Lu-NeoBOMB1, which had better therapeutic index for the organs receiving the highest doses.
Asunto(s)
Bombesina/química , Lutecio , Radioisótopos , Receptores de Bombesina/metabolismo , Animales , Bombesina/síntesis química , Bombesina/metabolismo , Humanos , Masculino , Ratones , Oligopéptidos/química , Tomografía de Emisión de Positrones , Neoplasias de la Próstata/diagnóstico por imagen , RadioquímicaRESUMEN
C-X-C chemokine receptor type 4 (CXCR4) is overexpressed in hematological and solid malignancies. LY2510924 is a potent peptide antagonist of CXCR4. A derivative of LY2510924, BL01, was evaluated for theranostic applications targeting CXCR4. Methods: BL01 was synthesized by solid phase approach. A Lys(ivDde) residue was added at the C-terminus of LY2510924 (cyclo[Phe-Tyr-Lys(iPr)-d-Arg-2-Nal-Gly-d-Glu]-Lys(iPr)-NH2). A DOTA chelator was conjugated to the side chain of the deprotected exogenous Lys residue. The binding affinity of Ga/Lu-BL01 was determined by competitive radioligand binding assays. BL01 was radiolabeled with 68GaCl3 or 177LuCl3. Biodistribution studies were performed in mice bearing Daudi Burkitt's lymphoma tumor xenografts at selected time points. PET imaging studies were performed with [68Ga]Ga-BL01, with blocking experiments performed with preinjection of LY2510924. The stability of [68Ga]Ga/[177Lu]Lu-BL01 was assessed in mouse plasma. Results: Ga-BL01 and Lu-BL01 have nanomolar affinity for CXCR4. [68Ga]Ga-BL01 was obtained in 58 ± 5% decay-corrected radiochemical yields and >99% radiochemical purity with a molar activity of 40 ± 11 GBq/µmol, while [177Lu]Lu-BL01 was obtained in 65 ± 6% decay-corrected radiochemical yields and >99% radiochemical purity with a molar activity of 120 ± 21 GBq/µmol. [68Ga]Ga-BL01 and [177Lu]Lu-BL01 were excreted primarily through the renal pathway. Daudi xenografts were clearly delineated in PET images with good contrast. On the basis of biodistribution data, tumor uptake of [68Ga]Ga-BL01 was 10.2 ± 2.56% injected dose per gram (%ID/g) at 1 h postinjection (p.i.). Spleen (12.6 ± 2.36 %ID/g) and lungs (13.2 ± 2.98 %ID/g), organs that express CXCR4, had high uptake as well. Preinjection of LY2510924 reduced average uptake of [68Ga]Ga-BL01 in tumors by 88%, demonstrating target specificity. The uptake of [68Ga]Ga-BL01 in tumor increased to 15.3 ± 1.86 %ID/g at 2 h p.i., with improved contrast. [177Lu]Lu-BL01 has similar pharmacokinetics as [68Ga]Ga-BL01 at 1 h p.i. The highest uptake was observed in tumor (14.0 ± 1.11 %ID/g), followed by the lungs (13.0 ± 1.27 %ID/g) and spleen (11.6 ± 1.78 %ID/g). The tumor uptake increased to 16.2 ± 2.69 %ID/g at 4 h p.i., before declining slightly to 10.1 ± 1.41 %ID/g at 24 h p.i. Both compounds were stable in vivo, as no metabolites were observed at 5 min p.i. Conclusions: [68Ga]Ga-BL01 and [177Lu]Lu-BL01 are a promising theranostic pair for imaging and endoradiotherapy of CXCR4-expressing malignancies.
Asunto(s)
Linfoma de Burkitt/radioterapia , Radioisótopos de Galio/uso terapéutico , Radiofármacos/uso terapéutico , Receptores CXCR4/metabolismo , Animales , Linfoma de Burkitt/metabolismo , Línea Celular , Radioisótopos de Galio/farmacocinética , Xenoinjertos , Pulmón/metabolismo , Pulmón/efectos de la radiación , Masculino , Ratones , Ratones Endogámicos NOD , Péptidos Cíclicos/farmacocinética , Péptidos Cíclicos/farmacología , Radioquímica/métodos , Radiofármacos/farmacocinética , Bazo/metabolismo , Bazo/efectos de la radiación , Nanomedicina Teranóstica/métodos , Distribución TisularRESUMEN
We designed and evaluated a novel albumin-binder-conjugated 177Lu-PSMA-617 derivative, 177Lu-HTK01169, with an extended blood retention time to maximize the radiation dose delivered to prostate tumors expressing prostate-specific membrane antigen (PSMA). PSMA-617 and HTK01169 that contained N-[4-( p-iodophenyl)butanoyl]-Glu as an albumin-binding motif were synthesized using the solid-phase approach. Binding affinity to PSMA was determined by in vitro competition-binding assay. 177Lu labeling was performed in acetate buffer (pH 4.5) at 90 °C for 15 min. SPECT/CT imaging, biodistribution, and endoradiotherapy studies were conducted in mice bearing PSMA-expressing LNCaP tumor xenografts. Radiation dosimetry was calculated using OLINDA software. Lu-PSMA-617 and Lu-HTK01169-bound PSMA with high affinity ( Ki values = 0.24 and 0.04 nM, respectively). SPECT imaging and biodistribution studies showed that 177Lu-PSMA-617 and 177Lu-HTK01169 were excreted mainly via the renal pathway. With fast blood clearance (0.68%ID/g at 1 h postinjection), the tumor uptake of 177Lu-PSMA-617 peaked at 1 h postinjection (15.1%ID/g) and gradually decreased to 7.91%ID/g at 120 h postinjection. With extended blood retention (16.6 and 2.10%ID/g at 1 and 24 h, respectively), the tumor uptake of 177Lu-HTK01169 peaked at 24 h postinjection (55.9%ID/g) and remained at the same level by the end of the study (120 h). Based on dosimetry calculations, 177Lu-HTK01169 delivered an 8.3-fold higher radiation dose than 177Lu-PSMA-617 to LNCaP tumor xenografts. For the endoradiotherapy study, the mice treated with 177Lu-PSMA-617 (18.5 MBq) all reached humane end point (tumor volume >1000 mm3) by Day 73 with a median survival of 58 days. Mice treated with 18.5, 9.3, 4.6, or 2.3 MBq of 177Lu-HTK01169 had a median survival of >120, 103, 61, and 28 days, respectively. With greatly enhanced tumor uptake and treatment efficacy compared to 177Lu-PSMA-617 in preclinical studies, 177Lu-HTK01169 warrants further investigation for endoradiotherapy of prostate cancer.
Asunto(s)
Antígenos de Superficie/metabolismo , Dipéptidos/administración & dosificación , Glutamato Carboxipeptidasa II/metabolismo , Compuestos Heterocíclicos con 1 Anillo/administración & dosificación , Lutecio/administración & dosificación , Neoplasias de la Próstata/radioterapia , Radioisótopos/administración & dosificación , Radiofármacos/administración & dosificación , Albúminas/metabolismo , Animales , Línea Celular Tumoral , Dipéptidos/química , Dipéptidos/farmacocinética , Glutamato Carboxipeptidasa II/antagonistas & inhibidores , Compuestos Heterocíclicos con 1 Anillo/química , Compuestos Heterocíclicos con 1 Anillo/farmacocinética , Humanos , Lutecio/química , Lutecio/farmacocinética , Masculino , Ratones , Antígeno Prostático Específico , Neoplasias de la Próstata/diagnóstico por imagen , Neoplasias de la Próstata/patología , Dominios y Motivos de Interacción de Proteínas , Radioisótopos/química , Radioisótopos/farmacocinética , Radiofármacos/química , Radiofármacos/farmacocinética , Tomografía Computarizada por Tomografía Computarizada de Emisión de Fotón Único/métodos , Distribución Tisular , Ensayos Antitumor por Modelo de XenoinjertoRESUMEN
BACKGROUND: Phantoms are commonly used to evaluate and compare the performance of imaging systems given the known ground truth. Positron emission tomography (PET) scanners are routinely validated using the NEMA image quality phantom, in which lesions are modeled using 10 to 37 mm fillable spheres. The NEMA phantom neglects, however, to model focal (3-10-mm), high-uptake lesions that are increasingly observed in prostate-specific membrane antigen (PSMA) PET images. PSMA-targeting radiopharmaceuticals allow for enhanced detection of metastatic prostate cancers. As such, there is significant need to develop an updated phantom which considers both the quantitative and lesion detectability of this new paradigm in oncological PET imaging. PURPOSE: In this work, we present the Quantitative PET Prostate Phantom (Q3P); a portable and modular phantom that can be used to improve and harmonize imaging protocols for 18F-PSMA PET scans. METHODS: A one-piece cylindrical phantom was designed effectively in two halves, which we call modules. Module 1 was designed to mimic lesions in the presence of background, and Module 2 mimicked very high contrast conditions (i.e., very low background) that can be observed in 18F-PSMA PET scans. Shell-less radioactive spheres (3-16-mm) were cast using epoxy resin mixed with sodium-22 (22Na), a long half-life positron emitter with positron range similar to 18F. To establish realistic lesion contrast, the 22Na spheres were mounted in a cylindrical chamber that can be filled with an 18F background (module 1). Thirteen exchangeable spherical cavity inserts (3-37-mm) were machined in two parts and solvent welded together, and filled with 18F (50 kBq/mL) to model lesions with very high contrast (module 2). Five 2.5-min PET scans were acquired on a 5-ring GE Discovery MI PET/CT scanner (General Electric, USA). Lesions were segmented using 41% of SUVmax fixed thresholding (41% FT) and recovery coefficients (RCs) were computed from 5 noise realizations. RESULTS: The manufactured phantom is portable (5.7 kg) and scan preparation takes less than 40 min. The total 22Na activity is 250 kBq, allowing it to be shipped as an exempt package under International Atomic Energy Agency (IAEA) regulations. Recovery coefficients, computed using PSF modeling and no post-reconstruction smoothing, were 130.3% (16 mm), 147.1% (10 mm), 87.2% (6 mm), and 7.0% (3 mm) for RCmax, which decreased to 91.1% (16 mm), 90.6% (10 mm), 53.2% (6 mm), and 3.6% (3 mm) for RCmean in the 22Na spheres. Comparatively, 18F sphere recovery was 110.7% (17 mm), 123.6% (10 mm), 106.5% (7 mm), and 23.3% (3 mm) for RCmax, which was reduced to 76.7% (17 mm), 77.7% (10 mm), 66.8% (7 mm), and 13.5% (3 mm), for RCmean. CONCLUSIONS: A standardized imaging phantom was developed for lesion quantification assessment in 18F-PSMA PET images. The phantom is configurable, providing users with the opportunity to modify background activity levels or sphere sizes according to clinical demands. Distributed to the community, the Q3P phantom has the potential to enable better assessment of lesion quantification and harmonization of 18F-PSMA PET imaging, which may lead to more robust predictive metrics and better outcome prediction in metastatic prostate cancer.
Asunto(s)
Radioisótopos de Flúor , Metástasis de la Neoplasia , Fantasmas de Imagen , Tomografía de Emisión de Positrones , Neoplasias de la Próstata , Masculino , Neoplasias de la Próstata/diagnóstico por imagen , Tomografía de Emisión de Positrones/instrumentación , Humanos , Procesamiento de Imagen Asistido por Computador/métodos , Garantía de la Calidad de Atención de Salud , Glutamato Carboxipeptidasa II/metabolismo , Control de CalidadRESUMEN
BACKGROUND AND OBJECTIVE: 177Lu-labeled prostate-specific membrane antigen (PSMA) radiopharmaceutical therapy (RPT) represents a pivotal advancement in addressing prostate cancer. However, existing therapies, while promising, remain incompletely understood and optimized. Computational models offer potential insights into RPTs, aiding in clinical drug delivery enhancement. In this study, we investigate the impact of various physiological parameters on the delivery of 177Lu-PSMA-617 RPT using the convection-diffusion-reaction (CDR) model. METHODS: Our investigation encompasses tumor geometry and surrounding tissue, characterized by well-defined boundaries and initial conditions. Utilizing the finite element method, we solve governing equations across a range of parameters: dissociation constant KD (1, 0.1, 0.01 [nM]), internalization rate (0.01-0.0001 [min-1]), diverse tumor shapes, and variable necrotic zone sizes. This model can provide an accurate analysis of radiopharmaceutical delivery from the injection site to the tumor cell, including drug transport in the vascular, interstitial, and intracellular spaces, and considering important parameters (e.g., drug extravasation from microvessels or to lymphatic vessels, the extracellular matrix, receptors, and intracellular space). RESULTS: Our findings reveal significant enhancements in tumor-absorbed doses as KD decreases. This outcome can be attributed to the higher affinity of radiopharmaceuticals for PSMA receptors as KD diminishes, facilitating a more efficient binding and retention of the therapeutic agent within the tumor microenvironment. Additionally, tumor-absorbed doses for KD â¼ 1 [nM] show an upward trend with higher internalization rates. This observation can be rationalized by considering that a greater internalization rate would result in a higher proportion of radiopharmaceuticals being taken up by tumor cells after binding to receptors on the cell surface. Notably, tumor shape and necrotic zone size exhibit limited influence on tumor absorbed dose. CONCLUSIONS: The present study employs the CDR model to explore the role of physiological parameters in shaping 177Lu-PSMA-617 RPT delivery. These findings provide insights for improving prostate cancer therapy by understanding radiopharmaceutical transport dynamics. This computational approach contributes to advancing our understanding of radiopharmaceutical delivery mechanisms and has implications for enhancing treatment efficacy.
Asunto(s)
Neoplasias de la Próstata , Radiofármacos , Masculino , Humanos , Radiofármacos/uso terapéutico , Radiofármacos/química , Neoplasias de la Próstata/diagnóstico por imagen , Neoplasias de la Próstata/radioterapia , Dipéptidos/uso terapéutico , Dipéptidos/química , Microambiente TumoralRESUMEN
OBJECTIVES: Accurate outcome prediction is important for making informed clinical decisions in cancer treatment. In this study, we assessed the feasibility of using changes in radiomic features over time (Delta radiomics: absolute and relative) following chemotherapy, to predict relapse/progression and time to progression (TTP) of primary mediastinal large B-cell lymphoma (PMBCL) patients. MATERIAL AND METHODS: Given the lack of standard staging PET scans until 2011, only 31 out of 103 PMBCL patients in our retrospective study had both pre-treatment and end-of-treatment (EoT) scans. Consequently, our radiomics analysis focused on these 31 patients who underwent [18F]FDG PET-CT scans before and after R-CHOP chemotherapy. Expert manual lesion segmentation was conducted on their scans for delta radiomics analysis, along with an additional 19 EoT scans, totaling 50 segmented scans for single time point analysis. Radiomics features (on PET and CT), along with maximum and mean standardized uptake values (SUVmax and SUVmean), total metabolic tumor volume (TMTV), tumor dissemination (Dmax), total lesion glycolysis (TLG), and the area under the curve of cumulative standardized uptake value-volume histogram (AUC-CSH) were calculated. We additionally applied longitudinal analysis using radial mean intensity (RIM) changes. For prediction of relapse/progression, we utilized the individual coefficient approximation for risk estimation (ICARE) and machine learning (ML) techniques (K-Nearest Neighbor (KNN), Linear Discriminant Analysis (LDA), and Random Forest (RF)) including sequential feature selection (SFS) following correlation analysis for feature selection. For TTP, ICARE and CoxNet approaches were utilized. In all models, we used nested cross-validation (CV) (with 10 outer folds and 5 repetitions, along with 5 inner folds and 20 repetitions) after balancing the dataset using Synthetic Minority Oversampling TEchnique (SMOTE). RESULTS: To predict relapse/progression using Delta radiomics between the baseline (staging) and EoT scans, the best performances in terms of accuracy and F1 score (F1 score is the harmonic mean of precision and recall, where precision is the ratio of true positives to the sum of true positives and false positives, and recall is the ratio of true positives to the sum of true positives and false negatives) were achieved with ICARE (accuracy = 0.81 ± 0.15, F1 = 0.77 ± 0.18), RF (accuracy = 0.89 ± 0.04, F1 = 0.87 ± 0.04), and LDA (accuracy = 0.89 ± 0.03, F1 = 0.89 ± 0.03), that are higher compared to the predictive power achieved by using only EoT radiomics features. For the second category of our analysis, TTP prediction, the best performer was CoxNet (LASSO feature selection) with c-index = 0.67 ± 0.06 when using baseline + Delta features (inclusion of both baseline and Delta features). The TTP results via Delta radiomics were comparable to the use of radiomics features extracted from EoT scans for TTP analysis (c-index = 0.68 ± 0.09) using CoxNet (with SFS). The performance of Deauville Score (DS) for TTP was c-index = 0.66 ± 0.09 for n = 50 and 0.67 ± 03 for n = 31 cases when using EoT scans with no significant differences compared to the radiomics signature from either EoT scans or baseline + Delta features (p-value> 0.05). CONCLUSION: This work demonstrates the potential of Delta radiomics and the importance of using EoT scans to predict progression and TTP from PMBCL [18F]FDG PET-CT scans.
RESUMEN
A comprehensive investigation of the Hg2+ coordination chemistry and 197m/gHg radiolabeling capabilities of cyclen-based commercial chelators, namely, DOTA and DOTAM (aka TCMC), along with their bifunctional counterparts, p-SCN-Bn-DOTA and p-SCN-Bn-TCMC, was conducted to assess the suitability of these frameworks as bifunctional chelators for the 197m/gHg2+ theranostic pair. Radiolabeling studies revealed that TCMC and DOTA exhibited low radiochemical yields (0%-6%), even when subjected to harsh conditions (80°C) and high ligand concentrations (10-4 M). In contrast, p-SCN-Bn-TCMC and p-SCN-Bn-DOTA demonstrated significantly higher 197m/gHg radiochemical yields (100% ± 0.0% and 70.9% ± 1.1%, respectively) under the same conditions. The [197 m/gHg]Hg-p-SCN-Bn-TCMC complex was kinetically inert when challenged against human serum and glutathione. To understand the differences in labeling between the commercial chelators and their bifunctional counterparts, non-radioactive natHg2+ complexes were assessed using NMR spectroscopy and DFT calculations. The NMR spectra of Hg-TCMC and Hg-p-SCN-Bn-TCMC suggested binding of the Hg2+ ion through the cyclen backbone framework. DFT studies indicated that binding of the Hg2+ ion within the backbone forms a thermodynamically stable product. However, competition can form between isothiocyanate binding and binding through the macrocycle, which was experimentally observed. The isothiocyanate bound coordination product was dominant at the radiochemical scale as, in comparison, the macrocycle bound product was seen at the NMR scale, agreeing with the DFT result. Furthermore, a bioconjugate of TCMC (TCMC-PSMA) targeting prostate-specific membrane antigen was synthesized and radiolabeled, resulting in an apparent molar activity of 0.089 MBq/nmol. However, the complex demonstrated significant degradation over 24 h when exposed to human serum and glutathione. Subsequently, cell binding assays were conducted, revealing a Ki value ranging from 19.0 to 19.6 nM. This research provides crucial insight into the effectiveness of current commercial chelators in the context of 197m/gHg2+ radiolabeling. It underscores the necessity for the development of specific and customized chelators to these unique "soft" radiometals to advance 197m/gHg2+ radiopharmaceuticals.
RESUMEN
Manual segmentation poses a time-consuming challenge for disease quantification, therapy evaluation, treatment planning, and outcome prediction. Convolutional neural networks (CNNs) hold promise in accurately identifying tumor locations and boundaries in PET scans. However, a major hurdle is the extensive amount of supervised and annotated data necessary for training. To overcome this limitation, this study explores semi-supervised approaches utilizing unlabeled data, specifically focusing on PET images of diffuse large B-cell lymphoma (DLBCL) and primary mediastinal large B-cell lymphoma (PMBCL) obtained from two centers. We considered 2-[18F]FDG PET images of 292 patients PMBCL (n = 104) and DLBCL (n = 188) (n = 232 for training and validation, and n = 60 for external testing). We harnessed classical wisdom embedded in traditional segmentation methods, such as the fuzzy clustering loss function (FCM), to tailor the training strategy for a 3D U-Net model, incorporating both supervised and unsupervised learning approaches. Various supervision levels were explored, including fully supervised methods with labeled FCM and unified focal/Dice loss, unsupervised methods with robust FCM (RFCM) and Mumford-Shah (MS) loss, and semi-supervised methods combining FCM with supervised Dice loss (MS + Dice) or labeled FCM (RFCM + FCM). The unified loss function yielded higher Dice scores (0.73 ± 0.11; 95% CI 0.67-0.8) than Dice loss (p value < 0.01). Among the semi-supervised approaches, RFCM + αFCM (α = 0.3) showed the best performance, with Dice score of 0.68 ± 0.10 (95% CI 0.45-0.77), outperforming MS + αDice for any supervision level (any α) (p < 0.01). Another semi-supervised approach with MS + αDice (α = 0.2) achieved Dice score of 0.59 ± 0.09 (95% CI 0.44-0.76) surpassing other supervision levels (p < 0.01). Given the time-consuming nature of manual delineations and the inconsistencies they may introduce, semi-supervised approaches hold promise for automating medical imaging segmentation workflows.
RESUMEN
PURPOSE: Automatic and accurate segmentation of lesions in images of metastatic castration-resistant prostate cancer has the potential to enable personalized radiopharmaceutical therapy and advanced treatment response monitoring. The aim of this study is to develop a convolutional neural networks-based framework for fully-automated detection and segmentation of metastatic prostate cancer lesions in whole-body PET/CT images. METHODS: 525 whole-body PET/CT images of patients with metastatic prostate cancer were available for the study, acquired with the [18F]DCFPyL radiotracer that targets prostate-specific membrane antigen (PSMA). U-Net (1)-based convolutional neural networks (CNNs) were trained to identify lesions on paired axial PET/CT slices. Baseline models were trained using batch-wise dice loss, as well as the proposed weighted batch-wise dice loss (wDice), and the lesion detection performance was quantified, with a particular emphasis on lesion size, intensity, and location. We used 418 images for model training, 30 for model validation, and 77 for model testing. In addition, we allowed our model to take n = 0,2, , 12 neighboring axial slices to examine how incorporating greater amounts of 3D context influences model performance. We selected the optimal number of neighboring axial slices that maximized the detection rate on the 30 validation images, and trained five neural networks with different architectures. RESULTS: Model performance was evaluated using the detection rate, Dice similarity coefficient (DSC) and sensitivity. We found that the proposed wDice loss significantly improved the lesion detection rate, lesion-wise DSC and lesion-wise sensitivity compared to the baseline, with corresponding average increases of 0.07 (p-value = 0.01), 0.03 (p-value = 0.01) and 0.04 (p-value = 0.01), respectively. The inclusion of the first two neighboring axial slices in the input likewise increased the detection rate by 0.17, lesion-wise DSC by 0.05, and lesion-wise mean sensitivity by 0.16. However, there was a minimal effect from including more distant neighboring slices. We ultimately chose to use a number of neighboring slices equal to 2 and the wDice loss function to train our final model. To evaluate the model's performance, we trained three models using identical hyperparameters on three different data splits. The results showed that, on average, the model was able to detect 80% of all testing lesions, with a detection rate of 93% for lesions with maximum standardized uptake values (SUVmax) greater than 5.0. In addition, the average median lesion-wise DSC was 0.51 and 0.60 for all the lesions and lesions with SUVmax>5.0, respectively, on the testing set. Four additional neural networks with different architectures were trained, and they both yielded stronger performance of segmenting lesions whose SUVmax>5.0 compared to the rest of lesions. CONCLUSION: Our results demonstrate that prostate cancer metastases in PSMA PET/CT images can be detected and segmented using CNNs. The segmentation performance strongly depends on the intensity, size, and the location of lesions, and can be improved by using specialized loss functions. Specifically, the models performed best in detection of lesions with SUVmax>5.0. Another challenge was to accurately segment lesions close to the bladder. Future work will focus on improving the detection of lesions with lower SUV values by designing custom loss functions that take into account the lesion intensity, using additional data augmentation techniques, and reducing the number of false lesions by developing methods to better separate signal from noise.
Asunto(s)
Tomografía Computarizada por Tomografía de Emisión de Positrones , Neoplasias de la Próstata , Masculino , Humanos , Tomografía Computarizada por Tomografía de Emisión de Positrones/métodos , Neoplasias de la Próstata/diagnóstico por imagen , Redes Neurales de la Computación , RadiofármacosRESUMEN
BACKGROUND: Potential risk associated with low-dose radiation exposures is often expressed using the effective dose (E) quantity. Other risk-related quantities have been proposed as alternatives. The recently introduced risk index (RI) shares similarities with E but expands the metric to incorporate medical imaging-appropriate risks factors including patient-specific size, age, and sex. PURPOSE: The aim of this work is to examine the RI metric for quantifying stochastic radiation risk and demonstrate its applications in nuclear imaging. The advantages in this improved metric may help the field progress toward stratified risk consideration in the course of patient management, improve efforts for procedure optimization, and support an evolution in the science of radiation risk assessment. METHODS: In this study we describe, implement, and calculate RI for various diagnostic nuclear imaging scenarios using reference biokinetics published in ICRP Publication 128 for commonly utilized radiopharmaceuticals. All absorbed dose, E and RI calculations were performed using the freely available MIRDcalc nuclear medicine dosimetry software; the organ specific risk parameters used in the software are also benchmarked in this text. The resulting RI and E values are compared and various trends in RI values identified. RESULTS: E and RI coefficients were calculated for 3016 use cases. Notably RI values vary depending on patient characteristics. Overall, across the population, global trends in RI values can be identified. In general, RI values were 2.15 times higher for females than males, due to higher risk coefficients and activities being distributed in smaller reference masses. The pediatric patients showed higher RIs than adults, as younger patients generally receive higher absorbed doses per administered activity, and are more radiosensitive, and have a longer projected lifespan at risk. A compendium of E and RI values is also provided in table format to serve as a reference for the community. CONCLUSIONS: RI is a rational quantity that could be used for justification, risk communication and protocol optimization in medical imaging. It has some advantages when compared to the long-utilized E value with respect to personalization, since accounts for patient size, age, sex, and natural incidence of cancer risk.
Asunto(s)
Radiometría , Radiofármacos , Masculino , Adulto , Femenino , Humanos , Niño , Dosis de Radiación , Radiometría/métodos , Programas Informáticos , Radiografía , Fantasmas de ImagenRESUMEN
High kidney and salivary gland uptake is a common feature of prostate-specific membrane antigen (PSMA)-targeted radioligands derived from the lysine-urea-glutamic acid (Lys-urea-Glu) pharmacophore. In this study we investigated if radioligands derived from lysine-urea-2-aminoadipic acid (Lys-urea-Aad), lysine-urea-S-carboxylmethylcysteine (Lys-urea-Cmc) and lysine-urea-O-carboxylmethylserine (Lys-urea-Cms) pharmacophores with/without an albumin binder could retain good PSMA-targeting capability but with minimized kidney and salivary gland uptake. Methods: HTK03177 and HTK03187 were obtained by replacing Aad in the previously reported Lys-urea-Aad-derived HTK03149 with Cmc and Cms, respectively. HTK03170, HTK04048 and HTK04028 were derived from HTK03149, HTK03177 and HTK03187, respectively, with the conjugation of an albumin-binding moiety, 4-(p-methoxyphenyl)butyric acid. In vitro competition binding assays were conducted using PSMA-expressing LNCaP prostate cancer cells and [18F]DCFPyL as the radioligand. Imaging and biodistribution studies of 68Ga-labeled HTK03177 and HTK03187, and 177Lu-labeled HTK03170, HTK04048 and HTK04028 were performed in LNCaP tumor-bearing mice. Radioligand therapy study of [177Lu]Lu-HTK03170 was carried out in LNCaP tumor-bearing mice and [177Lu]Lu-PSMA-617 was used for comparison. Results: The calculated Ki(PSMA) values of Ga-HTK03177, Ga-HTK03187, Lu-HTK03170, Lu-HTK04048 and Lu-HTK04028 were 5.0±2.4, 10.6±2.0, 1.6±0.4, 1.4±1.0 and 13.9±3.2 nM, respectively. PET Imaging and biodistribution studies at 1 h post-injection showed that both [68Ga]Ga-HTK03177 and [68Ga]Ga-HTK03187 had high uptake in LNCaP tumor xenografts (24.7±6.85 and 21.1±3.62 %ID/g, respectively) but minimal uptake in normal organs/tissues including kidneys (7.76±1.00 and 2.83±0.45 %ID/g, respectively) and salivary glands (0.22±0.02 and 0.16±0.02 %ID/g, respectively). SPECT imaging and biodistribution studies showed that the LNCaP tumor uptake of 177Lu-labeled HTK03170, HTK04048 and HTK04028 peaked at 4-24 h post-injection at ~43-65 %ID/g and was relatively sustained over time. Their peaked average uptake in kidneys (≤ 17.4 %ID/g) and salivary glands (≤ 2.92 %ID/g) was lower and continuously reduced over time. Radioligand therapy study showed that compared with [177Lu]Lu-PSMA-617 (37 MBq), a quarter dose of [177Lu]Lu-HTK03170 (9.3 MBq) led to a better median survival (63 vs 90 days). Conclusions: Our data demonstrate that that Lys-urea-Aad, Lys-urea-Cmc and Lys-urea-Cms are promising pharmacophores for the design of PSMA-targeted radioligands especially for radiotherapeutic applications to minimize toxicity to kidneys and salivary glands.
Asunto(s)
Radioisótopos de Galio , Lisina , Humanos , Masculino , Animales , Ratones , Farmacóforo , Distribución Tisular , Riñón , Albúminas , Ácido ButíricoRESUMEN
BACKGROUND: Dosimetry after radiopharmaceutical therapy with 177Lu (177Lu-RPT) relies on quantitative SPECT/CT imaging, for which suitable reconstruction protocols are required. In this study, we characterized for the first time the quantitative performance of a ring-shaped CZT-based camera using two different reconstruction algorithms: an ordered subset expectation maximization (OSEM) and a block sequential regularized expectation maximization (BSREM) combined with noise reduction regularization. This study lays the foundations for the definition of a reconstruction protocol enabling accurate dosimetry for patients treated with 177Lu-RPT. METHODS: A series of 177Lu-filled phantoms were acquired on a StarGuide™ (GE HealthCare), with energy and scatter windows centred at 208 (± 6%) keV and 185 (± 5%) keV, respectively. Images were reconstructed with the manufacturer implementations of OSEM (GE-OSEM) and BSREM (Q.Clear) algorithms, and various combinations of iterations and subsets. Additionally, the manufacturer-recommended Q.Clear-based reconstruction protocol was evaluated. Quantification accuracy, measured as the difference between the SPECT-based and the radionuclide calibrator-based activity, and noise were evaluated in a large cylinder. Recovery coefficients (RCs) and spatial resolution were assessed in a NEMA IEC phantom with sphere inserts. The reconstruction protocols considered suitable for clinical applications were tested on a cohort of patients treated with [177Lu]Lu-PSMA-I&T. RESULTS: The accuracy of the activity from the cylinder, although affected by septal penetration, was < 10% for all reconstructions. Both algorithms featured improved spatial resolution and higher RCs with increasing updates at the cost of noise build-up, but Q.Clear outperformed GE-OSEM in reducing noise accumulation. When the reconstruction parameters were carefully selected, similar values for noise (~0.15), spatial resolution (~1 cm) and RCs were found, irrespective of the reconstruction algorithm. Analogue results were found in patients. CONCLUSIONS: Accurate activity quantification is possible when imaging 177Lu with StarGuide™. However, the impact of septal penetration requires further investigations. GE-OSEM is a valid alternative to the recommended Q.Clear reconstruction algorithm, featuring comparable performances assessed on phantoms and patients.
RESUMEN
Theranostics is an emerging paradigm that combines imaging and therapy in order to personalize patient treatment. In nuclear medicine, this is achieved by using radiopharmaceuticals that target identical molecular targets for both imaging (using emitted gamma rays) and radiopharmaceutical therapy (using emitted beta, alpha or Auger-electron particles) for the treatment of various diseases, such as cancer. If the therapeutic radiopharmaceutical cannot be imaged quantitatively, a "theranostic pair" imaging surrogate can be used to predict the absorbed radiation doses from the therapeutic radiopharmaceutical. However, theranostic dosimetry assumes that the pharmacokinetics and biodistributions of both radiopharmaceuticals in the pair are identical or very similar, an assumption that still requires further validation for many theranostic pairs. In this review, we consider both same-element and different-element theranostic pairs and attempt to determine if factors exist which may cause inaccurate dose extrapolations in theranostic dosimetry, either intrinsic (e.g. chemical differences) or extrinsic (e.g. injecting different amounts of each radiopharmaceutical) to the radiopharmaceuticals. We discuss the basis behind theranostic dosimetry and present common theranostic pairs and their therapeutic applications in oncology. We investigate general factors that could create alterations in the behavior of the radiopharmaceuticals or the quantitative accuracy of imaging them. Finally, we attempt to determine if there is evidence showing some specific pairs as suitable for theranostic dosimetry. We show that there are a variety of intrinsic and extrinsic factors which can significantly alter the behavior among pairs of radiopharmaceuticals, even if they belong to the same chemical element. More research is needed to determine the impact of these factors on theranostic dosimetry estimates and on patient outcomes, and how to correctly account for them.
Asunto(s)
Neoplasias/terapia , Medicina Nuclear/métodos , Cintigrafía/métodos , Radiofármacos/administración & dosificación , Nanomedicina Teranóstica/métodos , Animales , HumanosRESUMEN
The ability of a computer to perform tasks normally requiring human intelligence or artificial intelligence (AI) is not new. However, until recently, practical applications in medical imaging were limited, especially in the clinic. With advances in theory, microelectronic circuits, and computer architecture as well as our ability to acquire and access large amounts of data, AI is becoming increasingly ubiquitous in medical imaging. Of particular interest to our community, radiomics tries to identify imaging features of specific pathology that can represent, for example, the texture or shape of a region in the image. This is conducted based on a review of mathematical patterns and pattern combinations. The difficulty is often finding sufficient data to span the spectrum of disease heterogeneity because many features change with pathology as well as over time and, among other issues, data acquisition is expensive. Although we are currently in the early days of the practical application of AI to medical imaging, research is ongoing to integrate imaging, molecular pathobiology, genetic make-up, and clinical manifestations to classify patients into subgroups for the purpose of precision medicine, or in other words, predicting a priori treatment response and outcome. Lung cancer is a functionally and morphologically heterogeneous disease. Positron emission tomography (PET) is an imaging technique with an important role in the precision medicine of patients with lung cancer that helps predict early response to therapy and guides the selection of appropriate treatment. Although still in its infancy, early results suggest that the use of AI in PET of lung cancer has promise for the detection, segmentation, and characterization of disease as well as for outcome prediction.
Asunto(s)
Inteligencia Artificial , Neoplasias Pulmonares , Humanos , Neoplasias Pulmonares/diagnóstico por imagen , Tomografía de Emisión de Positrones , Medicina de Precisión , PronósticoRESUMEN
BACKGROUND: Digital anthropomorphic phantoms, such as the 4D extended cardiac-torso (XCAT) phantom, are actively used to develop, optimize, and evaluate a variety of imaging applications, allowing for realistic patient modeling and knowledge of ground truth. The XCAT phantom defines the activity and attenuation for a simulated patient, which includes a complete set of organs, muscle, bone, and soft tissue, while also accounting for cardiac and respiratory motion. However, the XCAT phantom does not currently include the lymphatic system, critical for evaluating medical imaging tasks such as sentinel node detection, node density measurement, and radiation dosimetry. PURPOSE: In this study, we aimed to develop a scalable lymphatic system in the XCAT phantom, to facilitate improved research of the lymphatic system in medical imaging. Using this scalable lymphatic system, we modeled the lymph node conglomerate pathology that is characteristically observed in primary mediastinal B-cell lymphoma (PMBCL). As an extended application, we evaluated positron emission tomography (PET) image quantification of metabolic tumor volume (MTV) and total lesion glycolysis (TLG) of these simulated lymphomas, though the phantoms may be applied to other imaging modalities and study design paradigms (e.g., image quality, detection). METHODS: A template model for the lymphatic system was developed based on anatomical data from the Visible Human Project of the National Library of Medicine. The segmented nodes and vessels were fit with non-uniform rational basis spline surfaces, and multichannel large deformation diffeomorphic metric mapping was used to propagate the template to different XCAT anatomies. To model conglomerates observed in PMBCL, lymph nodes were enlarged, converged within the mediastinum, and tracer concentration was increased. We used the phantoms as inputs to a PET simulation tool, which generated images using ordered subsets expectation maximization reconstruction with 2-8 mm Gaussian filters. Fixed thresholding (FT) and gradient segmentation were used to determine MTV and TLG. Percent bias (%Bias) and coefficient of variation (COV) were computed as measures of accuracy and precision, respectively, for each MTV and TLG measurement. RESULTS: Using the methodology described above, we introduced a scalable lymphatic system in the XCAT phantom, which allows for the radioactivity and attenuation ground truth to be generated in 116 ± 2.5 s using a 2.3 GHz processor. Within the Rhinoceros interface, lymph node anatomy and function were modified to create a cohort of 10 phantoms with lymph node conglomerates. Using the lymphoma phantoms to evaluate PET quantification of MTV, mean %Bias values were -9.3%, -41.3%, and 20.9%, while COV values were 4.08%, 7.6%, and 3.4% using 25% FT, 40% FT, and gradient segmentations, respectively. Comparatively for TLG, mean %Bias values were -27.4%, -45.8%, and -16.0%, while COV values were 1.9%, 5.7%, and 1.4%, for the 25% FT, 40% FT, and gradient segmentations, respectively. CONCLUSIONS: In this work, we upgraded the XCAT phantom to include a lymphatic system, comprised of a network of 276 scalable lymph nodes and corresponding vessels. As an application, we created a cohort of phantoms with lymph node conglomerates to evaluate lymphoma quantification in PET imaging, which highlights an important application of this work.
Asunto(s)
Linfoma , Tomografía de Emisión de Positrones , Estados Unidos , Humanos , Sistema LinfáticoRESUMEN
Radiomics is an advanced image-processing framework, which extracts image features and considers them as biomarkers towards personalized medicine. Applications include disease detection, diagnosis, prognosis, and therapy response assessment/prediction. As radiation therapy aims for further individualized treatments, radiomics could play a critical role in various steps before, during and after treatment. Elucidation of the concept of radiomics-guided radiation therapy (RGRT) is the aim of this review, attempting to highlight opportunities and challenges underlying the use of radiomics to guide clinicians and physicists towards more effective radiation treatments. This work identifies the value of RGRT in various steps of radiotherapy from patient selection to follow-up, and subsequently provides recommendations to improve future radiotherapy using quantitative imaging features.
Asunto(s)
Procesamiento de Imagen Asistido por Computador , Oncología por Radiación , Humanos , Procesamiento de Imagen Asistido por Computador/métodos , Medicina de Precisión/métodosRESUMEN
The aim of this study was to investigate the effect of replacing Glu in the Lys-urea-Glu PSMA-targeting pharmacophore of [68Ga]Ga-HTK03041 with a close analog on the uptake of kidneys, salivary glands and PSMA-expressing tumor xenografts. Methods: HTK03161, HTK03149 and HTK03189A/B were obtained by replacing Glu in HTK03041 with Asp, Aad (L-2-aminoadipic acid) and Api (2-aminopimelic acid), respectively. PSMA binding affinities were measured by competition binding assays. PET imaging and biodistribution studies of 68Ga-labeled ligands were performed in LNCaP tumor-bearing mice. The best candidate HTK03149 was selected and radiolabeled with 177Lu, and SPECT imaging and biodistribution studies were performed in LNCaP tumor-bearing mice. Radiation dosimetry calculation was conducted using the OLINDA software. Radioligand therapy study was performed in LNCaP tumor-bearing mice treated with [177Lu]Lu-HTK03149 (9.3-148 MBq), [177Lu]Lu-PSMA-617 (37 MBq) or natLu-HTK03149 (500 pmol). Results: PSMA binding affinities (Ki) of Ga-HTK03161, Ga-HTK03149, Ga-HTK03189A and Lu-HTK03149 were 3.88±0.66, 6.99±0.80, 550±35.7 and 1.57±0.42 nM, respectively. PET imaging showed that all 68Ga-labeled HTK03161, HTK03149 and HTK03189A/B were excreted mainly via the renal pathway and had minimal uptake in all organs/tissues including kidneys and salivary glands. Tumor xenografts were clearly visualized in PET images of [68Ga]Ga-HTK03161 and [68Ga]Ga-HTK03149 but were barely visualized using [68Ga]Ga-HTK03189A/B. Tumor uptake values for [68Ga]Ga-HTK03161, [68Ga]Ga-HTK03149, [68Ga]Ga-HTK0189A and [68Ga]Ga-HTK03189B were 12.7±1.91, 19.1±6.37, 2.10±0.28 and 0.67±0.15 %IA/g, respectively at 1h post-injection, and their average kidney and salivary gland uptake values were 2.13-4.41 and 0.20-0.23 %IA/g, respectively. Longitudinal SPECT imaging studies showed that [177Lu]Lu-HTK03149 was excreted mainly through the renal pathway with high uptake in LNCaP tumors and minimal uptake in all normal organs/tissues. The tumor uptake of [177Lu]Lu-HTK03149 peaked at 4h post-injection (20.9±2.99 %IA/g) and the uptake was sustained over time. Compared to [177Lu]Lu-PSMA-617, [177Lu]Lu-HTK03149 had 145% increase in tumor absorbed dose but 70% less in kidney absorbed dose, leading to an 7.1-fold increase in tumor-to-kidney absorbed dose ratio. Radioligand therapy studies showed that only half of the [177Lu]Lu-PSMA-617 injected dosage was needed for [177Lu]Lu-HTK03149 to achieve the same median survival. Conclusion: Replacing Glu in the PSMA-targeting Lys-urea-Glu pharmacophore of [68Ga]Ga-HTK03041 with Asp and Aad generates [68Ga]Ga-HTK03161 and [68Ga]Ga-HTK03149, respectively, and the new derivatives retain high uptake in LNCaP tumors and have minimal uptake in normal organs/tissues including kidneys and salivary glands. [177Lu]Lu-HTK03149 also retain high uptake in LNCaP tumors and has minimal uptake in normal organs/tissues, and is, therefore, promising for clinical translation to treat prostate cancer.
Asunto(s)
Glutamato Carboxipeptidasa II , Neoplasias de la Próstata , Ácido 2-Aminoadípico/metabolismo , Animales , Antígenos de Superficie/metabolismo , Línea Celular Tumoral , Radioisótopos de Galio , Glutamato Carboxipeptidasa II/metabolismo , Humanos , Riñón/patología , Ligandos , Masculino , Ratones , Neoplasias de la Próstata/metabolismo , Radiofármacos , Glándulas Salivales , Distribución Tisular , Urea/metabolismoRESUMEN
PURPOSE: Mantle cell lymphoma (MCL) is associated with poor survival. The purpose of this study was to assess whether the C-X-C chemokine receptor type 4 (CXCR4) is a useful target for imaging and radioligand therapy of MCL, using a novel pair of radioligands, [68Ga]Ga and [177Lu]Lu-BL02. EXPERIMENTAL DESIGN: We performed a retrospective analysis of 146 patients with MCL to evaluate CXCR4 expression and its correlation with outcomes. Guided by in silico methods, we designed BL02, a new radioligand labelled with 68Ga or 177Lu for PET imaging and therapy, respectively. We performed imaging and biodistribution studies in xenograft models with varying CXCR4 expression. We evaluated [177Lu]Lu-BL02 in MCL models, and evaluated its potential for therapy in Z138 MCL xenografts. RESULTS: Phosphorylated and nonphosphorylated CXCR4 expression were correlated with poor survival in patients with MCL and characterized by unique underlying molecular signatures. [68Ga]Ga-BL02 uptake correlated with CXCR4 expression, and localized lesions in a metastatic xenograft model. [177Lu]Lu-BL02 showed high uptake in MCL xenografts. Therapy studies with a single dose in the Z138 model showed tumor regression and improved survival compared with a control group. Upon regrowth, the treated mice experienced concurrent metastasis alongside localized xenograft regrowth, and recurrent lesions showed enhanced CXCR4 signaling. CONCLUSIONS: CXCR4 is an independent factor of poor prognosis for MCL and a promising target for imaging and radioligand therapy. [68Ga]Ga-BL02 showed high contrast to visualize CXCR4-expressing xenografts for PET imaging and [177Lu]Lu-BL02 induced rapid tumor regression in a preclinical model of MCL.
Asunto(s)
Linfoma de Células del Manto , Adulto , Animales , Línea Celular Tumoral , Humanos , Linfoma de Células del Manto/diagnóstico por imagen , Linfoma de Células del Manto/tratamiento farmacológico , Linfoma de Células del Manto/radioterapia , Ratones , Tomografía de Emisión de Positrones/métodos , Receptores CXCR4/genética , Receptores CXCR4/metabolismo , Estudios Retrospectivos , Distribución TisularRESUMEN
This article is the second part in our machine learning series. Part 1 provided a general overview of machine learning in nuclear medicine. Part 2 focuses on neural networks. We start with an example illustrating how neural networks work and a discussion of potential applications. Recognizing that there is a spectrum of applications, we focus on recent publications in the areas of image reconstruction, low-dose PET, disease detection, and models used for diagnosis and outcome prediction. Finally, since the way machine learning algorithms are reported in the literature is extremely variable, we conclude with a call to arms regarding the need for standardized reporting of design and outcome metrics and we propose a basic checklist our community might follow going forward.