Distinguishing Progression from Pseudoprogression in Glioblastoma Using 18F-Fluciclovine PET

Rationale: Accurate differentiation between tumor progression (TP) and pseudoprogression remains a critical unmet need in neuro-oncology. ¹⁸F-fluciclovine is a widely available synthetic amino acid PET radiotracer. In this study, we aimed to assess the value of ¹⁸F-fluciclovine PET for differentiating pseudoprogression from TP in a prospective cohort of patients with suspected radiographic recurrence of glioblastoma. Methods: We enrolled 30 glioblastoma patients with radiographic progression after first-line chemoradiotherapy who were planned for surgical resection. Patients underwent pre-operative ¹⁸F-fluciclovine PET and MRI. Relative percentages of viable tumor and therapy-related changes observed in histopathology were quantified and categorized as TP (≥50% viable tumor), mixed TP (<50% and >10% viable tumor), or pseudoprogression (≤10% viable tumor). Results: Eighteen patients had TP, 4 mixed TP, and 8 pseudoprogression. Patients with TP/mixed TP had significantly higher 40-50 minutes SUV_max (6.64+ 1.88 vs 4.11± 1.52, P = 0.009) compared to patients with pseudoprogression. A 40-50 minutes SUV_max cut-off of 4.66 provided 90% sensitivity and 83% specificity for differentiation of TP/mixed TP from pseudoprogression (Area under the curve (AUC)=0.86). Relative cerebral blood volume (rCBVmax) cut-off 3.672 provided 90% sensitivity and 71% specificity for differentiation of TP/mixed TP from Pseudoprogression (AUC=0.779). Combining a 40-50 minutes SUV_max cut-off of 4.66 and a rCBVmax cut-off of 3.67 on MRI provided 100% sensitivity and 80% specificity for differentiating TP/mixed TP from Pseudoprogression (AUC=0.95). Conclusion: ¹⁸F-fluciclovine PET uptake can accurately differentiate pseudoprogression from TP in glioblastoma, with even greater accuracy when combined with multi-parametric MRI. Given the wide availability of ¹⁸F-fluciclovine, larger, multicenter studies are warranted to determine whether amino acid PET with ¹⁸F-fluciclovine should be used in the routine assessment of post-treatment glioblastoma.

Federated learning enables big data for rare cancer boundary detection

Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.

NIMG-45. DISTINGUISHING PROGRESSION FROM PSEUDOPROGRESSION IN GLIOBLASTOMA: COMBINED USE OF 18F-FLUCICLOVINE PET AND MULTI-PARAMETRIC MRI

PURPOSE

Differentiation of tumor progression (TP) from pseudoprogression (PsP) is a major unmet need in post-treatment glioblastoma (GBM). 18F-Fluciclovine is a synthetic amino acid PET radiotracer with higher uptake in tumor tissue vs. areas of treatment-related change. We investigated the value of 18F-Fluciclovine PET for differentiating PsP from TP independent from and in combination with multi-parametric MRI.

METHODS

We prospectively enrolled 30 patients with GBM with a new or enlarging contrast-enhancing lesion on MRI after chemoradiotherapy who were planned for surgical resection of the lesion. Patients underwent pre-operative 18F-Fluciclovine PET and multi-parametric MRI. Following surgery, the relative percentages of viable tumor and therapy-related changes observed in histopathology were quantified. Patients were categorized as TP if viable tumor represented ≥ 50% of the specimen, mixed TP if &lt; 50% and &gt; 10%, and PsP if ≤ 10%.

RESULTS

18 patients had TP, 4 had mixed TP, and 8 PsP. Patients with TP/mixed TP had a significantly higher 40-50 minutes SUVmax (6.64 + 1.88 vs 4.11± 1.52, p=0.009) and an SUVmax cut-off of 4.66 provided 90% sensitivity and 83% specificity for differentiation of TP/mixed TP from PsP (AUC=0.856). A maximum cerebral blood volume (CBVmax) cut-off of 3.67 provided 90% sensitivity and 71% specificity for differentiation of TP/mixed TP from PsP (AUC=0.779). Combining a 40-50 minutes SUVmax cut-off of 4.662 and a relative CBVmax cut-off of 3.67 provided 100% sensitivity and 80% specificity for differentiating TP/mixed TP from PsP (AUC=0.95). The time activity curve patterns and time to peaks were not different between the groups. Normalization of PET parameters to normal brain parenchyma were not helpful to differentiate the groups due to variability in radiotracer uptake in normal brain between subjects.

CONCLUSION

18F-Fluciclovine PET uptake can accurately differentiate PsP from TP in GBM patients, with even more accurate differentiation achieved when combined with MRI.

JS07.5.A 18F-fluciclovine PET and multi-parametric MRI to distinguish pseudoprogression from tumor progression in post-treatment glioblastoma

Background

Differentiation of tumor progression (TP) from pseudoprogression (PsP) is a major unmet need in glioblastoma (GBM). 18F-Fluciclovine is a synthetic amino acid PET radiotracer with higher uptake in tumor tissue vs. areas of treatment-related change. We aimed to assess the combined value of 18F-Fluciclovine PET and multi-parametric MRI for differentiating PsP from TP.

Material and Methods

We enrolled 30 patients with GBM with a new or enlarging contrast-enhancing lesion on MRI after chemoradiotherapy who were planned for surgical resection of the lesion. Patients underwent pre-operative 18F-Fluciclovine PET and multi-parametric MRI. Following surgery, the relative percentages of viable tumor and therapy-related changes observed on histopathology were quantified. Patients were categorized as TP if viable tumor represented ≥ 50% of the specimen, mixed TP if &lt; 50% and &gt; 10%, and PsP if ≤ 10%. SUVmax, SUVpeak, and 50% threshold SUVmean were calculated and normalized to contralateral brain, pituitary gland, and superior sagittal sinus (SSS). The least absolute shrinkage and selection operator (LASSO) was used to determine the variables most predictive of tumor percentage. The strength of association between the primary outcome and selected variables was assessed by Pearson’s or Point-biserial correlation.

Results

18 patients with TP, 4 with mixed TP-PsP, and 8 with PsP were included. There was a positive correlation between 50% threshold SUV mean measured from PET images acquired 50-60 minutes post-injection and rCBVmax by MRI and tumor percentage by histology (r= 0.56; p= 0.004 and r=0.50; p=0.012 respectively). 40-50 minutes SUVmax (OR=1.78 rpb=0.51) and rCBVmax (OR=1.64, rpb=0.48) were positively correlated with tumor TP/mixed TP group. Patients who demonstrated TP/mixed TP-PsP had significantly higher 40-50 minutes SUVmax compared to patients with histological PsP (6.71±2.03 vs 3.93±1.63; p=0.012). 40-50 minutes SUVmax cut-off of 4.46 provided 90% sensitivity and 80% specificity for differentiation of TP/mixed TP-PsP from PsP (AUC=0.88). Combining a 40-50 minutes SUVmax cut-off of 4.46 and an rCBVmax cut-off of 3.67 provided 100% sensitivity and 100% specificity for differentiating TP/mixed TP-PsP from PsP (AUC=1). Patients who demonstrated TP had a significantly higher 40-50 minutes SUVmax compared to patients with histological PsP (6.99±2.06 vs 3.93±1.63; p=0.008). A 40-50 minutes SUVmax cut-off of 4.66 provided 94% sensitivity and 80% specificity for differentiation of TP from PsP (AUC=0.89).

Conclusion

18F-Fluciclovine PET uptake is positively correlated with viable tumor quantification on histology and can accurately differentiate PsP from TP in patients with GBM. Further independent studies are required to cross-validate these promising early results.

NIMG-55. AUGMENTED INTELLIGENCE IS SUPERIOR TO ARTIFICIAL INTELLIGENCE! HUMAN-COMPUTER SYNERGY FOR GENERATING HIGH QUALITY GLIOBLASTOMA SUB-REGION SEGMENTATIONS

PURPOSE

Artificial intelligence (AI) is poised to improve diagnostic methods in neuro-oncologic imaging and contribute to patient management by analyzing pre-operative MRI scans. AI results are better interpreted by compartmentalizing glioblastoma into distinct sub-regions, i.e., necrotic core, enhancing tumor, peritumoral T2/FLAIR signal abnormality (ED). Manual delineation of these sub-regions by expert neuroradiologists is impractical, requiring hours for intricate cases. Computer-aided segmentation (CAS) can mitigate this issue but is limited in the quality of the produced segmentations. We hypothesize that CAS followed by expert refinements is more practical/time-efficient.

METHODS

CAS was used on a total of 359 glioblastoma patients with four MRI sequences (T1, T1Gd, T2, T2-FLAIR) from each patient. All segmentations were sent to expert neuroradiologist annotators for manual refinements. Once refined, our team including two senior attending neuroradiologists with ≥13 years of experience each, reviewed and either approved or returned the segmentations to individual annotators for further refinements. Total time required to refine and review the finalized segmentations was measured.

RESULTS

Following one round of refinements by expert annotators, 244/359 (68%) segmentations were approved by our team while 115/359 (32%) segmentations contained a variety of errors that required a second round of refinements. The most common observed errors were 1) missed ED in the anterior/inferior temporal lobes and corpus callosum (37/115 cases, 32%) and 2) erroneous segmentation of normal choroid plexus and blood vessels (14/115 cases, 12%). The expert annotators required 120 hours to refine all 359 segmentations, and our team required 26 additional hours to review them, resulting in 24 minutes/segmentation following CAS.

CONCLUSION

Our findings support the value of a well-communicated annotation protocol to coordinate CAS and expert annotators. With CAS, our team and expert annotators rapidly finalized segmentations for 359 glioblastoma patients, demonstrating the value of a synergistic approach to creating high quality tumor sub-region segmentations.

Quantifying Perfusion Properties with DCE-MRI Using a Dictionary Matching Approach

Perfusion properties can be estimated from pharmacokinetic models applied to DCE-MRI data using curve fitting algorithms; however, these suffer from drawbacks including the local minimum problem and substantial computational time. Here, a dictionary matching approach is proposed as an alternative. Curve fitting and dictionary matching were applied to simulated data using the dual-input single-compartment model with known perfusion property values and 5 in vivo DCE-MRI datasets. In simulation at SNR 60 dB, the dictionary estimate had a mean percent error of 0.4-1.0% for arterial fraction, 0.5-1.4% for distribution volume, and 0.0% for mean transit time. The curve fitting estimate had a mean percent error of 1.1-2.1% for arterial fraction, 0.5-1.3% for distribution volume, and 0.2-1.8% for mean transit time. In vivo, dictionary matching and curve fitting showed no statistically significant differences in any of the perfusion property measurements in any of the 10 ROIs between the methods. In vivo, the dictionary method performed over 140-fold faster than curve fitting, obtaining whole volume perfusion maps in just over 10 s. This study establishes the feasibility of using a dictionary matching approach as a new and faster way of estimating perfusion properties from pharmacokinetic models in DCE-MRI.

Repeatability and reproducibility of 3D MR fingerprinting relaxometry measurements in normal breast tissue

BACKGROUND

The 3D breast magnetic resonance fingerprinting (MRF) technique enables T₁ and T₂ mapping in breast tissues. Combined repeatability and reproducibility studies on breast T₁ and T₂ relaxometry are lacking.

PURPOSE

To assess test-retest and two-visit repeatability and interscanner reproducibility of the 3D breast MRF technique in a single-institution setting.

STUDY TYPE

Prospective.

SUBJECTS

Eighteen women (median age 29 years, range, 22-33 years) underwent Visit 1 scans on scanner 1. Ten of these women underwent test-retest scan repositioning after a 10-minute interval. Thirteen women had Visit 2 scans within 7-15 days in same menstrual cycle. The remaining five women had Visit 2 scans in the same menstrual phase in next menstrual cycle. Five women were also scanned on scanner 2 at both visits for interscanner reproducibility.

FIELD STRENGTH/SEQUENCE

Two 3T MR scanners with the 3D breast MRF technique.

ASSESSMENT

T₁ and T₂ MRF maps of both breasts.

STATISTICAL TESTS

Mean T₁ and T₂ values for normal fibroglandular tissues were quantified at all scans. For variability, between and within-subjects coefficients of variation (bCV and wCV, respectively) were assessed. Repeatability was assessed with Bland-Altman analysis and coefficient of repeatability (CR). Reproducibility was assessed with interscanner coefficient of variation (CoV) and Wilcoxon signed-rank test.

RESULTS

The bCV at test-retest scans was 9-12% for T₁ , 7-17% for T₂ , wCV was <4% for T₁ , and <7% for T₂ . For two visits in same menstrual cycle, bCV was 10-15% for T₁ , 13-17% for T₂ , wCV was <7% for T₁ and <5% for T₂ . For two visits in the same menstrual phase, bCV was 6-14% for T₁ , 15-18% for T₂ , wCV was <7% for T₁ , and <9% for T₂ . For test-retest scans, CR for T₁ and T₂ were 130 msec and 11 msec. For two visit scans, CR was <290 msec for T₁ and 10-14 msec for T₂ . Interscanner CoV was 3.3-3.6% for T₁ and 5.1-6.6% for T₂ , with no differences between interscanner measurements (P = 1.00 for T₁ , P = 0.344 for T₂ ).

DATA CONCLUSION

3D breast MRF measurements are repeatable across scan timings and scanners and may be useful in clinical applications in breast imaging.

LEVEL OF EVIDENCE

2 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2019;50:1133-1143.

Free-Breathing 3D Liver Perfusion Quantification Using a Dual-Input Two-Compartment Model

The purpose of this study is to test the feasibility of applying a dual-input two-compartment liver perfusion model to patients with different pathologies. A total of 7 healthy subjects and 11 patients with focal liver lesions, including 6 patients with metastatic adenocarcinoma and 5 with hepatocellular carcinoma (HCC), were examined. Liver perfusion values were measured from both focal liver lesions and cirrhotic tissues (from the 5 HCC patients). Compared to results from volunteer livers, significantly higher arterial fraction, fractional volume of the interstitial space, and lower permeability-surface area product were observed for metastatic lesions, and significantly higher arterial fraction and lower vascular transit time were observed for HCCs (P < 0.05). Significantly lower arterial fraction and higher vascular transit time, fractional volume of the vascular space, and fractional volume of the interstitial space were observed for metastases in comparison to HCCs (P < 0.05). For cirrhotic livers, a significantly lower total perfusion, lower fractional volume of the vascular space, higher fractional volume of the interstitial space, and lower permeability-surface area product were noted in comparison to volunteer livers (P < 0.05). Our findings support the possibility of using this model with 3D free-breathing acquisitions for lesion and diffuse liver disease characterization.