Hypermetabolic Axillary Lymphadenopathy After Shingrix Vaccination for Varicella Zoster Virus.

ABSTRACT: A 67-year-old woman underwent staging 18 F-FDG PET/CT scan for recently diagnosed breast cancer. Her scan showed a highly hypermetabolic right breast mass, with ipsilateral hypermetabolic axillary lymph nodes. The contralateral axillary lymph nodes were also enlarged with avid FDG uptake, alongside focal increased uptake in the left deltoid muscle. Upon investigation, the patient reported receiving the new zoster recombinant adjuvanted varicella zoster vaccine (Shingrix, GlaxoSmithKline) 4 days before the scan. The lymph node uptake could be potential pitfall for cancer staging.

Virtual reality-based training and pre-operative planning for head and neck sentinel lymph node biopsy.

OBJECTIVE: Sentinel lymph node biopsy (SLNB) is crucial for managing head and neck skin cancer. However, variable lymphatic drainage can complicate SLN detection when using Single-Photon Emission Computed Tomography (SPECT) or lymphoscintigraphy. Virtual Reality (VR) can contribute to pre-operative planning by simulating a realistic 3D model, which improves orientation. VR can also facilitate real-patient training outside the operating room. This study explored using a VR platform for pre-operative planning in head and neck skin cancer patients undergoing SLNBs and assessed its value for residential training. MATERIALS AND METHODS: In this prospective technology pilot study, attending surgeons and residents who performed 21 SLNB operations on patients with head and neck skin cancers (81% males, mean age 69.2 ± 11.3) used a VR simulation model based on each patient's pre-operative SPECT scan to examine patient-specific anatomy. After surgery, they completed a questionnaire on the efficiency of the VR simulation as a pre-operative planning tool and training device for residents. RESULTS: The attending surgeons rated the VR model's accuracy at 8.3 ± 1.6 out of 10. Three-quarters (76%) of residents reported increased confidence after using VR. The physicians rated the platform's contribution to residents' training at 7.4 ± 2.1 to 8.9 ± 1.3 out of 10. CONCLUSION: A VR SLNB simulation can accurately portray marked sentinel lymph nodes. It was rated high as a surgical planning and teaching tool among attending surgeons and residents alike and may play a role in pre-operative planning and resident training. Further studies are needed to explore its applications in practice.

The Role of 68 Ga-FAPI PET/CT in Breast Cancer Response Assessment and Follow-Up.

PURPOSE: 68 Ga-fibroblast activation protein inhibitor (FAPI), a new PET/CT radiotracer targeting cancer-associated fibroblasts in tumor microenvironment, can detect many types of cancer. We aimed to assess whether it can also be used for response assessment and follow-up. METHODS: We followed up patients with FAPI-avid invasive lobular breast cancer (ILC) before and after treatment changes and correlated qualitative maximal intensity projection images and quantitative tumor volume with CT results and blood tumor biomarkers. RESULTS: Six consenting ILC breast cancer patients (53 ± 8 years old) underwent a total of 24 scans (baseline for each patient and 2-4 follow-up scans). We found a strong correlation between 68 Ga-FAPI tumor volume and blood biomarkers ( r = 0.7, P < 0.01), but weak correlation between CT and 68 Ga-FAPI maximal intensity projection-based qualitative response assessment. CONCLUSIONS: We found a strong correlation between ILC progression and regression (as assessed by blood biomarkers) and 68 Ga-FAPI tumor volume. 68 Ga-FAPI PET/CT could possibly be used for disease response assessment and follow-up.

Combination of FDG PET/CT Radiomics and Clinical Parameters for Outcome Prediction in Patients with Hodgkin's Lymphoma.

PURPOSE: The aim of the study is to evaluate the prognostic value of a joint evaluation of PET and CT radiomics combined with standard clinical parameters in patients with HL. METHODS: Overall, 88 patients (42 female and 46 male) with a median age of 43.3 (range 21-85 years) were included. Textural analysis of the PET/CT images was performed using freely available software (LIFE X). 65 radiomic features (RF) were evaluated. Univariate and multivariate models were used to determine the value of clinical characteristics and FDG PET/CT radiomics in outcome prediction. In addition, a binary logistic regression model was used to determine potential predictors for radiotherapy treatment and odds ratios (OR), with 95% confidence intervals (CI) reported. Features relevant to survival outcomes were assessed using Cox proportional hazards to calculate hazard ratios with 95% CI. RESULTS: albumin (p = 0.034) + ALP (p = 0.028) + CT radiomic feature GLRLM GLNU mean (p = 0.012) (Area under the curve (AUC): 95% CI (86.9; 100.0)-Brier score: 3.9, 95% CI (0.1; 7.8) remained significant independent predictors for PFS outcome. PET-SHAPE Sphericity (p = 0.033); CT grey-level zone length matrix with high gray-level zone emphasis (GLZLM SZHGE mean (p = 0.028)); PARAMS XSpatial Resampling (p = 0.0091) as well as hemoglobin results (p = 0.016) remained as independent factors in the final model for a binary outcome as predictors of the need for radiotherapy (AUC = 0.79). CONCLUSION: We evaluated the value of baseline clinical parameters as well as combined PET and CT radiomics in HL patients for survival and the prediction of the need for radiotherapy treatment. We found that different combinations of all three factors/features were independently predictive of the here evaluated endpoints.

The Role of 68 Ga-FAPI PET/CT in Detection of Metastatic Lobular Breast Cancer.

PURPOSE: Invasive lobular breast cancer (ILC) may be hard to detect using conventional imaging modalities and usually shows less avidity to 18 F-FDG PET/CT. 68 Ga-fibroblast activation protein inhibitor (FAPI) PET/CT has shown promising results in detecting non- 18 F-FDG-avid cancers. We aimed to assess the feasibility of detecting metastatic disease in patients with non- 18 F-FDG-avid ILC. METHODS: This prospective study included patients with metastatic ILC, infiltrative to soft tissues, which was not 18 F-FDG avid. The patients underwent 68 Ga-FAPI PET/CT for evaluation, which was correlated with the fully diagnostic CT performed at the same time. RESULTS: Seven women (aged 57 ± 10 years) were included. Among the 30 organs and structures found to be involved by tumor, the number of findings observed by FAPI PET/CT was significantly higher than that observed by CT alone ( P = 0.022), especially in infiltrative soft tissue and serosal locations. CONCLUSIONS: This small pilot trial suggests a role for 68 Ga-FAPI PET/CT in ILC, which needs to be confirmed by subsequent trials.

FDG PET/CT radiomics as a tool to differentiate between reactive axillary lymphadenopathy following COVID-19 vaccination and metastatic breast cancer axillary lymphadenopathy: a pilot study.

OBJECTIVES: To evaluate if radiomics with machine learning can differentiate between F-18-fluorodeoxyglucose (FDG)-avid breast cancer metastatic lymphadenopathy and FDG-avid COVID-19 mRNA vaccine-related axillary lymphadenopathy. MATERIALS AND METHODS: We retrospectively analyzed FDG-positive, pathology-proven, metastatic axillary lymph nodes in 53 breast cancer patients who had PET/CT for follow-up or staging, and FDG-positive axillary lymph nodes in 46 patients who were vaccinated with the COVID-19 mRNA vaccine. Radiomics features (110 features classified into 7 groups) were extracted from all segmented lymph nodes. Analysis was performed on PET, CT, and combined PET/CT inputs. Lymph nodes were randomly assigned to a training (n = 132) and validation cohort (n = 33) by 5-fold cross-validation. K-nearest neighbors (KNN) and random forest (RF) machine learning models were used. Performance was evaluated using an area under the receiver-operator characteristic curve (AUC-ROC) score. RESULTS: Axillary lymph nodes from breast cancer patients (n = 85) and COVID-19-vaccinated individuals (n = 80) were analyzed. Analysis of first-order features showed statistically significant differences (p < 0.05) in all combined PET/CT features, most PET features, and half of the CT features. The KNN model showed the best performance score for combined PET/CT and PET input with 0.98 (± 0.03) and 0.88 (± 0.07) validation AUC, and 96% (± 4%) and 85% (± 9%) validation accuracy, respectively. The RF model showed the best result for CT input with 0.96 (± 0.04) validation AUC and 90% (± 6%) validation accuracy. CONCLUSION: Radiomics features can differentiate between FDG-avid breast cancer metastatic and FDG-avid COVID-19 vaccine-related axillary lymphadenopathy. Such a model may have a role in differentiating benign nodes from malignant ones. KEY POINTS: • Patients who were vaccinated with the COVID-19 mRNA vaccine have shown FDG-avid reactive axillary lymph nodes in PET-CT scans. • We evaluated if radiomics and machine learning can distinguish between FDG-avid metastatic axillary lymphadenopathy in breast cancer patients and FDG-avid reactive axillary lymph nodes. • Combined PET and CT radiomics data showed good test AUC (0.98) for distinguishing between metastatic axillary lymphadenopathy and post-COVID-19 vaccine-associated axillary lymphadenopathy. Therefore, the use of radiomics may have a role in differentiating between benign from malignant FDG-avid nodes.

Risk stratification for relapsed/refractory classical Hodgkin lymphoma integrating pretransplant Deauville score and residual metabolic tumor volume.

Pretransplant Deauville score (DS) is an imaging biomarker used for risk stratification in relapsed/refractory classical Hodgkin lymphoma (cHL). However, the prognostic value of residual metabolic tumor volume (rMTV) in patients with DS 4-5 has been less well characterized. We retrospectively assessed 106 patients with relapsed/refractory cHL who underwent autologous stem cell transplantation. Pretransplant DS was determined as 1-3 (59%) and 4-5 (41%), with a markedly inferior event-free survival (EFS) in patients with DS 4-5 (hazard ratio [HR], 3.14; p = .002). High rMTV41% (rMTVhigh , ≥4.4 cm3 ) predicted significantly poorer EFS in patients with DS 4-5 (HR, 3.70; p = .014). In a multivariable analysis, we identified two independent factors predicting treatment failure: pretransplant DS combined with rMTV41% and disease status (primary refractory vs. relapsed). These two factors allow to stratify patients into three groups with divergent 2-year EFS: 89% for low-risk (51%; relapsed disease and either pretransplant DS 1-3 or DS 4-5/rMTVlow ; HR 1), 65% for intermediate-risk (28%; refractory disease and either DS 1-3 or DS 4-5/rMTVlow ; HR 3.26), and 45% for high-risk (21%; DS 4-5/rMTVhigh irrespective of disease status; HR 7.61) groups. Pretransplant DS/rMTV41% combination and disease status predict the risk of post-transplant treatment failure and will guide risk-stratified approaches in relapsed/refractory cHL.

Imaging of Hematological Patients in the Era of COVID-19.

The COVID-19 pandemic, caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), resulted in changes in management and imaging routines for patients with hematological malignancies. Treating physicians had to familiarize themselves with a new disease, with distinct imaging manifestations, sometimes overlapping with other infections prevalent in this patient population. In some aspects, infected hematological patients might exhibit a different disease course, and routine imaging in asymptomatic hematological patients may result in unexpected COVID-19 findings, implying covert infection, that should be further explored. Furthermore, some complications of hematological diseases and treatments may present with findings similar to COVID-19 manifestations, and treating physicians must consider both possibilities in the differential diagnosis. In this review, we aimed to present the influence the COVID-19 pandemic had on hematological malignancy imaging.

The Utility of 68Ga-PSMA PET/CT in Decisions Regarding Administering Salvage Radiotherapy to Men with Prostate Cancer.

BACKGROUND: Numerous papers have described 68Ga-prostate-specific membrane antigen (PSMA) positron emission tomography/computed tomography (PET/CT)'s sensitivity in identifying prostate cancer (PCa) recurrence. This study aimed to characterize the role of 68Ga-PSMA PET/CT in deciding to re-irradiate pelvic structures. METHODS: 68Ga-PSMA PET/CT scans performed at Sheba Medical Center over seven years in 113 men were reviewed. All had undergone radiation to the prostate (70, 61.9%) or post-radical prostatectomy radiation to the prostate fossa (PF) (43, 48.1%), and had local or oligometastatic PCa recurrence and received salvage radiotherapy (SRT) based on PET/CT findings. RESULTS: Mean age was 70.7 years. The mean grade group was 2.9; the mean prostate-specific antigen was 9.0. The 68Ga-PSMA PET/CT positive findings included: 37 (32.7%) in the prostate, 23 (20.4%) in seminal vesicles, 7 (6.2%) in the PF, and 3 (2.7%) in the seminal vesicle fossa. The mean standardized uptake value was 10.6 ± 10.2 (range: 1.4-61.6); the mean lesion size was 1.8 ± 3.5 mm (range: 0.5-5.1). SRT was directed toward the prostate and seminal vesicles in 48 (42.5%), PF in 18 (15.9%), and intrapelvic lymph node and bone in 47 (41.6%). Toxicities were mostly mild to moderate. CONCLUSION: 68Ga-PSMA PET/CT-identified relapse with targeted SRT was well-tolerated and may result in less onerous treatments.

COVID-19 mRNA Vaccination: Age and Immune Status and Its Association with Axillary Lymph Node PET/CT Uptake.

With hundreds of millions of coronavirus disease 2019 (COVID-19) messenger RNA (mRNA)-based vaccine doses planned to be delivered worldwide in the upcoming months, it is important to recognize PET/CT findings in recently vaccinated immunocompetent or immunocompromised patients. We aimed to assess PET/CT uptake in the deltoid muscle and axillary lymph nodes of patients who received a COVID-19 mRNA-based vaccine and to evaluate its association with patient age and immune status. Methods: All consecutive adults who underwent PET/CT scans with any radiotracer at our center during the first month of a national COVID-19 vaccination rollout (between December 23, 2020, and January 27, 2021) and had received the vaccination were included. Data on clinical status, laterality, and time from vaccination were prospectively collected, retrospectively analyzed, and correlated with deltoid muscle and axillary lymph node uptake. Results: Of 426 eligible subjects (median age, 67 ± 12 y; 49% female), 377 (88%) underwent PET/CT with 18F-FDG, and positive axillary lymph node uptake was seen in 45% of them. Multivariate logistic regression analysis revealed a strong inverse association between positive 18F-FDG uptake in ipsilateral lymph nodes and patient age (odds ratio [OR], 0.57; 95% CI, 0.45-0.72; P < 0.001), immunosuppressive treatment (OR, 0.37; 95% CI, 0.20-0.64; P = 0.003), and presence of hematologic disease (OR, 0.44; 95% CI, 0.24-0.8; P = 0.021). No such association was found for deltoid muscle uptake. The number of days from the last vaccination and the number of vaccine doses were also significantly associated with increased odds of positive lymph node uptake. Conclusion: After mRNA-based COVID-19 vaccination, a high proportion of patients showed ipsilateral lymph node axillary uptake, which was more common in immunocompetent patients. This information will help with the recognition of PET/CT pitfalls and may hint about the patient's immune response to the vaccine.

The association between lesion tracer uptake on 68Ga-DOTATATE PET with morphological response to 177Lu-DOTATATE therapy in patients with progressive metastatic neuroendocrine tumors.

PURPOSE: To determine in a group of patients with progressive metastatic neuroendocrine tumors (PM-NETs) treated with 177Lu-DOTATATE whether a correlation exists between somatostatin receptor (SSTR)-2 expression in various tumors on baseline 68Ga-DOTATATE PET and their response to therapy. A secondary aim was to determine whether an association exists between tumor product of diameter (POD) and PET-derived Krenning score. MATERIALS METHODS: Patients treated PM-NETs who had SSTR-2 overexpression (SSTR-RADS 5) on screening 68Ga-DOTATATE PET and CT at baseline and 3 months after therapy completion were included. Marker lesions on baseline CT were reassessed on CT after therapy using adapted Southwest Oncology Group solid tumor evaluation criteria. For each lesion, bidimensional diameter on CT and SSTR expression on PET (SSTR-RADS uptake score & PET-derived Krenning score) were recorded. Logistic regression models fitted through generalized estimating equations were used to assess for an association between SSTR expression and response to therapy, or lesion's POD. RESULTS: Forty-one patients with SSTR-RADS 5 PM-NETs treated with 177Lu-DOTATATE were included. There were 135 marker lesions (mean 3.2 lesions/patient) with Krenning score of 4 (n = 74), 3 (n = 44) or 2 (n = 17). There was no association found between SSTR-2 expression, as determined by SSTR-RADS uptake score or PET-derived Krenning score, and POD or response to therapy. CONCLUSION: In patients with SSTR-RADS 5 PM-NETs treated with 177Lu-DOTATATE, there was similar response to therapy for all lesions with PET-generated Krenning score ≥2. No correlation was found between lesion's POD and level of tracer uptake.

PSMA PET/CT to evaluate response to SBRT for prostate cancer bone metastases.

BACKGROUND: In the current study we evaluated 68Ga PSMA PET/ CT to measure local control of bone metastasis in oligometastatic prostate cancer patients treated with SBRT. MATERIALS AND METHODS: After the institutional review board approval, a retrospective review of medical records of consecutive prostate cancer patients treated between 2014 and 2018 was conducted. Only medical records of patients that were treated with SBRT for bone metastasis and had pre-and post-SBRT 68Ga PSMA PET/CT scans were included in our study. Data extracted from the medical files included patient-related (age), disease-related (Gleason score, site of metastasis), and treatment-related factors and outcomes. RESULTS: During the study period, a total of 12 patients (15 lesions) were included, with a median age of 73 years. The median follow-up was 26.5 months (range 13-45 months). Median time of 68Ga PSMA PET/ CT follow up was 17.0 months (range 3-39 months). The median pre-treatment PSA was 2 ng/mL (range 0.56-44 ng/mL) vs. post treatment PSA nadir of 0.01 ng/mL (0.01-4.32) with a median time to nadir of 7 months (range, 2-12). Local control was 93% during the follow up period and there was correlation with PS MA avidity on PE T. None patients developed recurrences in the treated bone. None of the patients had grade 3 or more toxicities during follow-up. CONCLUSIONS: SBRT is a highly effective and safe method for treatment of prostate cancer bone metastases. More studies are required to determine if SBRT provides greater clinical benefit than standard fractionation for oligometastatic prostate cancer patients. 68Ga PSMA PET/CT should be further investigated for delineation and follow-up.

Detection of clinically significant prostate cancer with 18F-DCFPyL PET/multiparametric MR.

PURPOSE: To assess whether 18F-DCFPyL PET/multiparametric (mp)MR contributes to the diagnosis of clinically significant (cs) prostate cancer (PCa) compared to mpMR in patients with suspicion of PCa, or patients being considered for focal ablative therapies (FT). PATIENTS AND METHODS: This ethics review board-approved, prospective study included 55 men with suspicion of PCa and negative systematic biopsies or clinically discordant low-risk PCa (n = 21) or those being considered for FT (n = 34) who received 18F-DCFPyL PET/mpMR. Each modality, PET, mpMR, and PET/MR (using the PROMISE classification), was assessed independently. All suspicious lesions underwent PET/MR-ultrasound fusion biopsies. RESULTS: There were 45/55 patients (81.8%) that had histologically proven PCa and 41/55 (74.5%) were diagnosed with csPCa. Overall, 61/114 lesions (53.5%) identified on any modality were malignant; 49/61 lesions (80.3%) were csPCa. On lesion-level analysis, for detection of csPCa, the sensitivity of PET was higher than that of mpMR and PET/MR (86% vs 67% and 69% [p = 0.027 and 0.041, respectively]), but at a lower specificity (32% vs 85% and 86%, respectively [p < 0.001]). The performance of MR and PET/MR was comparable. For identification of csPCa in PI-RADS ≥ 3 lesions, the AUC (95% CI) for PET, mpMR, and PET/MR was 0.75 (0.65-0.86), 0.69 (0.56-0.82), and 0.78 (0.67-0.89), respectively. The AUC for PET/MR was significantly larger than that of mpMR (p = 0.04). CONCLUSION: PSMA PET detects more csPCa than mpMR, but at low specificity. The performance PET/MR is better than mpMR for detection of csPCa in PI-RADS ≥ 3 lesions. CLINICAL REGISTRATION: NCT03149861.

Physiologic and hypermetabolic breast 18-F FDG uptake on PET/CT during lactation.

OBJECTIVE: To investigate the patterns of breast cancer-related and lactation-related 18F-FDG uptake in breasts of lactating patients with pregnancy-associated breast cancer (PABC) and without breast cancer. METHODS: 18F-FDG-PET/CT datasets of 16 lactating patients with PABC and 16 non-breast cancer lactating patients (controls) were retrospectively evaluated. Uptake was assessed in the tumor and non-affected lactating tissue of the PABC group, and in healthy lactating breasts of the control group, using maximum and mean standardized uptake values (SUVmax and SUVmean, respectively), and breast-SUVmax/liver-SUVmean ratio. Statistical tests were used to evaluate differences and correlations between the groups. RESULTS: Physiological uptake in non-breast cancer lactating patients' breasts was characteristically high regardless of active malignancy status other than breast cancer (SUVmax = 5.0 ± 1.7, n = 32 breasts). Uptake correlated highly between the two breasts (r = 0.61, p = 0.01), but was not correlated with age or lactation duration (p = 0.24 and p = 0.61, respectively). Among PABC patients, the tumors demonstrated high 18F-FDG uptake (SUVmax = 7.8 ± 7.2, n = 16), which was 326-643% higher than the mostly low physiological FDG uptake observed in the non-affected lactating parenchyma of these patients (SUVmax = 2.1 ± 1.1). Overall, 18F-FDG uptake in lactating breasts of PABC patients was significantly decreased by 59% (p < 0.0001) compared with that of lactating controls without breast cancer. CONCLUSION: 18F-FDG uptake in lactating tissue of PABC patients is markedly lower compared with the characteristically high physiological uptake among lactating patients without breast cancer. Consequently, breast tumors visualized by 18F-FDG uptake in PET/CT were comfortably depicted on top of the background 18F-FDG uptake in lactating tissue of PABC patients. KEY POINTS: • FDG uptake in the breast is characteristically high among lactating patients regardless of the presence of an active malignancy other than breast cancer. • FDG uptake in non-affected lactating breast tissue is significantly lower among PABC patients compared with that in lactating women who do not have breast cancer. • In pregnancy-associated breast cancer patients, 18F-FDG uptake is markedly increased in the breast tumor compared with uptake in the non-affected lactating tissue, enabling its prompt visualization on PET/CT.

Fecal microbiota transplant promotes response in immunotherapy-refractory melanoma patients.

The gut microbiome has been shown to influence the response of tumors to anti-PD-1 (programmed cell death-1) immunotherapy in preclinical mouse models and observational patient cohorts. However, modulation of gut microbiota in cancer patients has not been investigated in clinical trials. In this study, we performed a phase 1 clinical trial to assess the safety and feasibility of fecal microbiota transplantation (FMT) and reinduction of anti-PD-1 immunotherapy in 10 patients with anti-PD-1-refractory metastatic melanoma. We observed clinical responses in three patients, including two partial responses and one complete response. Notably, treatment with FMT was associated with favorable changes in immune cell infiltrates and gene expression profiles in both the gut lamina propria and the tumor microenvironment. These early findings have implications for modulating the gut microbiota in cancer treatment.

A risk model for relapsed/refractory aggressive NHL integrating clinical risk factors and pretransplant Deauville score.

There are limited data regarding the combined value of the pretransplant Deauville score (DS) from a positron emission tomography scan and clinical risk factors in patients with relapsed/refractory aggressive non-Hodgkin lymphoma (NHL). We performed a retrospective analysis to assess the prognostic role of pretransplant DS in patients with relapsed/refractory aggressive NHL who underwent salvage chemotherapy and autologous stem cell transplantation (ASCT). We identified 174 eligible patients between January 2013 and March 2019. In multivariable analysis, pretransplant DS, B symptoms, and secondary International Prognostic Index (sIPI) were independent risk factors for event-free survival (EFS). These variables were used to derive an integrated risk score that categorized 166 patients with available information for all risk factors into 3 groups: low (n = 92; 55.4%), intermediate (n = 48; 28.9%), and high (n = 26; 15.7%). The new prognostic index showed a strong association with EFS (low-risk vs intermediate-risk hazard ratio [HR], 3.94; 95% confidence interval [CI], 2.16-7.17; P < .001; low-risk vs high-risk HR, 10.83; 95% CI, 5.81-20.19; P < .001) and outperformed models based on clinical risk factors or DS alone. These results were validated in 60 patients from an independent external cohort (low-risk vs intermediate-risk HR, 4.04; 95% CI, 1.51-10.82; P = .005; low-risk vs high-risk HR, 10.49; 95% CI, 4.11-26.73; P < .001). We propose and validate a new prognostic index that risk-stratifies patients undergoing salvage chemotherapy followed by ASCT, thereby identifying patients at high risk for posttransplant treatment failure.

Convolutional neural networks for improving image quality with noisy PET data.

GOAL: PET is a relatively noisy process compared to other imaging modalities, and sparsity of acquisition data leads to noise in the images. Recent work has focused on machine learning techniques to improve PET images, and this study investigates a deep learning approach to improve the quality of reconstructed image volumes through denoising by a 3D convolution neural network. Potential improvements were evaluated within a clinical context by physician performance in a reading task. METHODS: A wide range of controlled noise levels was emulated from a set of chest PET data in patients with lung cancer, and a convolutional neural network was trained to denoise the reconstructed images using the full-count reconstructions as the ground truth. The benefits, over conventional Gaussian smoothing, were quantified across all noise levels by observer performance in an image ranking and lesion detection task. RESULTS: The CNN-denoised images were generally ranked by the physicians equal to or better than the Gaussian-smoothed images for all count levels, with the largest effects observed in the lowest-count image sets. For the CNN-denoised images, overall lesion contrast recovery was 60% and 90% at the 1 and 20 million count levels, respectively. Notwithstanding the reduced lesion contrast recovery in noisy data, the CNN-denoised images also yielded better lesion detectability in low count levels. For example, at 1 million true counts, the average true positive detection rate was around 40% for the CNN-denoised images and 30% for the smoothed images. CONCLUSION: Significant improvements were found for CNN-denoising for very noisy images, and to some degree for all noise levels. The technique presented here offered however limited benefit for detection performance for images at the count levels routinely encountered in the clinic.