Predicting FDG-PET Images From Multi-Contrast MRI Using Deep Learning in Patients With Brain Neoplasms

BACKGROUND

18 F-fluorodeoxyglucose (FDG) positron emission tomography (PET) is valuable for determining presence of viable tumor, but is limited by geographical restrictions, radiation exposure, and high cost.

PURPOSE

To generate diagnostic-quality PET equivalent imaging for patients with brain neoplasms by deep learning with multi-contrast MRI.

STUDY TYPE

Retrospective.

SUBJECTS

Patients (59 studies from 51 subjects; age 56 ± 13 years; 29 males) who underwent 18 F-FDG PET and MRI for determining recurrent brain tumor.

FIELD STRENGTH/SEQUENCE

3T; 3D GRE T1, 3D GRE T1c, 3D FSE T2-FLAIR, and 3D FSE ASL, 18 F-FDG PET imaging.

ASSESSMENT

Convolutional neural networks were trained using four MRIs as inputs and acquired FDG PET images as output. The agreement between the acquired and synthesized PET was evaluated by quality metrics and Bland-Altman plots for standardized uptake value ratio. Three physicians scored image quality on a 5-point scale, with score ≥3 as high-quality. They assessed the lesions on a 5-point scale, which was binarized to analyze diagnostic consistency of the synthesized PET compared to the acquired PET.

STATISTICAL TESTS

The agreement in ratings between the acquired and synthesized PET were tested with Gwet's AC and exact Bowker test of symmetry. Agreement of the readers was assessed by Gwet's AC. P = 0.05 was used as the cutoff for statistical significance.

RESULTS

The synthesized PET visually resembled the acquired PET and showed significant improvement in quality metrics (+21.7% on PSNR, +22.2% on SSIM, -31.8% on RSME) compared with ASL. A total of 49.7% of the synthesized PET were considered as high-quality compared to 73.4% of the acquired PET which was statistically significant, but with distinct variability between readers. For the positive/negative lesion assessment, the synthesized PET had an accuracy of 87% but had a tendency to overcall.

CONCLUSION

The proposed deep learning model has the potential of synthesizing diagnostic quality FDG PET images without the use of radiotracers.

EVIDENCE LEVEL

3 TECHNICAL EFFICACY: Stage 2.

OUP accepted manuscript

Background

Peptide receptor radionuclide therapy (PRRT) with radiolabeled somatostatin receptor (SSR) analogs is now an established systemic treatment for neuroendocrine tumors (NET). However, more short- and long-term data about renal and hepatotoxicity is needed. Here we present our experience in this clinical scenario.

Methods

Eighty-six patients with progressive SSR-expressing malignancies underwent PRRT with Lu-177 Dotatate and were followed up for up to 2 years. Laboratory tests were done 1 week before each cycle and every 2 months at follow-up. Hepatic and renal toxicity was determined based on NCI CTCAE V5.0.

Results

55/86 (64%) patients completed all 4 cycles of PRRT; 18/86 (20.9%) are currently being treated; 13/86 (15.1%) had to discontinue PRRT: 4/13 (31%) due to hematologic toxicity, 9/13 (69%) due to non-PRRT-related comorbidities. Out of the patients who finished treatment, only transient grade 2 toxicities were observed during PRRT: hypoalbuminemia in 5.5% (3/55), and renal toxicity (serum creatinine and estimated glomerular filtration rate) in 1.8% (1/55). No grade 3 or 4 liver and renal toxicity occurred. Patients presenting with impaired liver or renal function prior to PRRT, either improved or had stable findings. No deterioration was observed.

Conclusion

Peptide receptor radionuclide therapy does not have a negative impact on liver and renal function, even in patients with pre-existing impaired parameters. No grade 3 or 4 hepatic or renal toxicity was identified. Only transient grade 2 hypoalbuminemia in 5.5% and nephrotoxicity in 1.8% of patients were seen during PRRT.

Single institution experience with peptide receptor radionuclide therapy (PRRT) in neuroendocrine tumors (NET)

623

Background: Neuroendocrine tumors (NETs) are rare but increasing in incidence. The only curative treatment is surgery, which in many cases is not an option due to metastatic disease at diagnosis. The NETTER-1 study showed high efficacy and low toxicity of peptide receptor radionuclide therapy (PRRT) for midgut NETs. Here, we present our initial experience with PRRT in the treatment of patients with NET. Methods: Fifty-two patients (27 males and 25 females; 37 - 81 yo, mean ± SD: 61.8 ± 10.6 years) with documented progressive NET (25 pancreas, 17 small intestine, 1 coecum, 4 unknown primary, 3 paragangliomas, and 2 pheochromocytomas) were referred to undergo PRRT at our institution from July 2018 to September 2019. Laboratory tests were obtained at baseline, 1 week before each cycle and every 3 months following treatment. Progression-free survival (PFS), objective response rate (ORR) and toxicity were assessed. An interim overall survival (OS) analysis was performed. Results, when possible, were compared with the NETTER-1 trial. Lines of therapy were documented. Results: 22/52 (42%) patients completed all 4 cycles of PRRT. 18/52 (34%) patients are currently being treated. 12/52 (23%) patients had to discontinue treatment. Hematotoxicity was the only side effect which can be related to PRRT. The 6-month and 9-month PFS rate was 82.4% and 66.8% respectively vs. 89% and 84% in the NETTER-1 trial. The ORR was 36% vs. 18% in the NETTER-1 trial. In the interim OS analysis, 6 deaths occurred. In contrast to the NETTER-1 study, PRRT in our patient cohort was performed later in the course of treatment (median lines of therapy before PRRT = 4 ±1.3 (range 1-6)). Conclusions: Our preliminary data show overall good results of PRRT in patients with NETs. However, compared to the NETTER-1 trial, PFS is shorter which is most likely due to the extensive pretreatment, but ORR was higher. [Table: see text]