Table 2.
Overview of Included AA PET Studies
| Study | Tracer and Study Type | Number of Patients | Chemotherapy Agents Used in Recurrence | OCEBM Level of Evidence | Measurements | Results and Conclusions |
|---|---|---|---|---|---|---|
| Hutterer et al., 201113 | Static FET Retrospective |
11 | BEV + IR | 4 | PFS OS (rOS) SUV Tumor reduction |
FET-PET detected BEV-treatment failure earlier than RANO criteria with MRI alone, with a mean time benefit of 9 weeks - AUC-ROC curve for PFS using tumor volume reduction was 0.875 ± 0.117 with a sensitivity of 87.5% and specificity of 100% - FET responders had longer PFS than RANO responders, but not significantly - FET volumes decreased in long-term survivors (PFS ≥ 6 months) but increased in short-term survivors (PFS< 6 months) - Contrast-enhancement on MRI does not always correspond to areas of metabolically active tumor with FET-PET |
| Harris et al., 201228 | Static F-DOPA Retrospective |
24 | BEV +/- IR | 4 | PFS OS SUV TBR (PRM) |
F-DOPA-PET stratified patients by short-term and long-term survival - Analyzed PRMs corresponding to regions of pretreatment contrast enhancement on MRI - PRMs with increased F-DOPA uptake stratified patients by 3-month PFS and 6-month OS, with increased uptake corresponding to shorter PFS and OS (sensitivity of 75%, specificity of 70%) - When comparing 2 post-BEV treatment PET scans, an increased volume fraction of F-DOPA uptake stratified patients by short- and long-term PFS and OS (sensitivity of 91%, specificity of 83%) |
| Galldiks et al., 201311 | Static and Dynamic FET Prospective |
10 | BEV + IR | 3 | PFSOS (tOS) SUV TBR Tumor reduction TAC TTP |
Static and dynamic FET predicted BEV-treatment failure earlier than what was reported by Hutterer et al., with a median of 4.9 weeks and a mean time benefit of 10.5 weeks - Changes in TBRmax greater than 17% differentiated responders (PFS greater than 6 mos) from nonresponders with a sensitivity of 83% and specificity of 100% - Response on FET had sig longer PFS than nonresponders (9 vs 3 mos) and OS (23 mos vs 3.5 mos) - Mean FET tumor volumes decreased in long term survivors but unchanged in short term survivors Differentiated responders from nonresponders by characterizing TTP and kinetic patterns - TTP at baseline could differentiate responders and nonresponders with an AUC of 1.0 and sensitivity and specificity of 100% - Shorter TTP at baseline and follow-up in nonresponders - Nonresponders more likely to have an early peak of FET uptake than a constant descent (Type 3 kinetic pattern) Proposed BEV-response is suggested if 2 of 3 criteria are met: 1. Increase of TTP ≥ 10 min at follow-up 2. TTP ≥ 25 min at baseline 3. Type 1 (SUV peaks at the end) or 2 kinetics (SUV peaks at least half-way through and plateaus or has slow descent) |
| Wardak et al., 201430 | Dynamic F-DOPA Prospective |
21 | BEV + IR | 3 | PFS OS SUV TAC |
Looking at dynamic parameters can provide predictive information of OS after BEV - The average PFS and OS was 183 days and 13.2 months, respectively - OS and PFS significantly correlated (r = 0.85, P < .0001) - FLT more predictive of OS than F-DOPA (Multiple Linear Regression = R2 = 0.83 for both, R2 = 0.82 for FLT alone) |
| Schwarzenberg et al., 201420 | Static F-DOPA Prospective |
30 | BEV +/- IR | 3 | PFS OS SUV TBR MTV |
F-DOPA PET at 2 weeks was statistically significant of distinguishing treatment response to antiangiogenic therapy with the PET parameter MTV - F-DOPA PET SUV parameters were not predictive of outcomes - MTV was highly prognostic and predictive of treatment response, most significantly at 2 weeks with an OS (P < .0001) and PFS (P = .001) - ROC curve identified 18 mL as the optimal threshold for MTV |
| Garcia et al., 201512 | Static MET Case report |
1 | BEV | NA | SUV | Pseudoprogression evident on MRI after BEV - MRI demonstrated increased lesion size after initiation of BEV-MET-PET demonstrated normalization of metabolic activity |
| Beppu et al., 20167 | Static MET Pilot study |
20 | BEV + TMZ | 3 | PFS SUV |
MET-PET at 8 weeks differentiated pseudoresponders from true responders after BEV and is correlated with PFS - MRI-response at 4 and 8 weeks is associated with longer PFS - 20% of MRI-responders were reclassified as pseudoresponse by MET-PET at 8 weeks - MET-PET alone and in addition to MRI at 8 weeks predicts longer PFS, but not at 4 weeks - Lower SUVT/N at 8 weeks versus baseline was associated with longer PFS than increased SUVT/N |
| Fleischmann et al., 201725 | Static and Dynamic FETRetrospective | 57a | BEV | 4 | OS (PRS) SUV TBR BTV TAC TTP |
TTPmin in dynamic FET-PET may be used for prognosis in rHGG treated with radiation and BEV - Earlier TTPmin is associated with shorter PRS (P = .027) - Time from pretreatment PET to initiation of radiation and BEV significantly affected TTPmin - Suggest rHGG with decreasing TAC may be evaluated using TTPmin for spatial characterization of aggressive tumor regions |
| Deuschl et al., 201726 | Static C-MET Prospective | 11 | BEV +/- LOM | 3 | PFS OS SUVT/N SUVMax MTV |
Suggested a method for determining treatment response to BEV using PET/MRI: if T/N ratio on PET is greater than 1.6, then response is determined by decrease in T/N ratio of 25%. If T/N is less than 1.6, use RANO guidelines - MRI: 0/11 Complete response, 6/11 partial response, 4/11 stable disease, 1/11 progressive disease - PET: 4/6 partial responders on MRI were metabolic responders on C-MET PET; 2/4 of stable disease on MRI were responders on C-MET PET; 4 PET nonresponders including 2 cases of pseudoresponse (MRI and PET discordant) - PFS and OS was predicted by reduction of T/N of greater than 25% on C-MET and the RANO classification of tumor response on MRI |
| Galldiks et al., 201824 | Static F-FET Prospective | 21 | BEV + LOM | 3 | TBR MTV OS |
Both relative change in MTV and absolute MTV on F-FET at follow-up after BEV predicted long term survival >9 months. - Patients were considered responders to BEV/LOM if they survived >9 months - Median OS of responders versus nonresponders was 12 and 6 months, respectively - Response on FET-PET significantly predicted OS greater than 9 months (P < .05), however, MRI results at 10 weeks did not predict an OS greater than 9 months (P = .203) - Reduction of TBRmax of FET-PET by 27% or a reduction of TBR by 17% distinguished responders from nonresponders. - Patients with MTV less than 5 mL survived significantly longer with average OS of 11 months than those who had a larger tumor volume on follow-up—median of 6 months (P < .001). |
| George et al., 201823 | Static F-FET Prospective |
11 | BEV | 3 | PFS OS SUV (VOI) TBR (Voxel-wise Spearman correlation coefficient) |
Concluded that there is moderate correlation between FET PET uptake and MRI contrast enhancement after BEV - PFS and OS after BEV was 111 days and 223 days, respectively - The voxel-wise Spearman correlation coefficient between FET uptake and T2/FLAIR signal intensity was 0.65 before BEV and 0.61 after BEV (P = .256). - Increased correlation between PET uptake and MRI enhancement after BEV was associated with lower PFS (P < .001) and OS (P = .049). - Post-treatment to pretreatment PET uptake ratio greater than 0.7 was associated with lower PFS and OS. |
| Bashir et al., 201922 | Static F-FET Retrospective |
50a | BEV | 4 | OS TBR BTV |
F-FET PET can distinguish post-treatment changes to BEV from tumor recurrence at 6-month follow-up, and predictive of OS - No significant difference found with F-FET uptakes between measurable and nonmeasurable lesions (TBRmax, 3.0 vs 3.2; TBRmean, 2.0 vs 2.0; and BTV 12 cm3 vs 10.5 cm3; P > .05) - ROC curve threshold of 2.0 for TBRmax, 1.8 for TBRmean, and 0.55 cm3 for BTV in differentiating between recurrence or treatment changes with a sensitivity of 99% and specificity of 94% (P < .0001) - F-FET parameters are higher in patients with recurrent glioblastoma versus post-treatment changes (TBRmax, 3.2 vs 1.6; TBR mean, 2.0 vs 1.6; and BTV 14.8 cm3 vs 0.01 cm3; P < .0001) - Increased BTV correlated with lower OS - 99% accuracy to distinguish recurrence from late post-treatment changes |
| Beppu et al., 201927 | Static C-MET Prospective | 24 | BEV + TMZ | 3 | PFS SUV T/N |
SUVT/N obtained with C-MET PET was more predictive of PFS in patients with rHGG treated with BEV at 8 weeks compared to T/N found using average relative cerebral blood flow from arterial spin labeling - Median PFS was 123 days - Changes in T/N ratios and tumor volumes on ASL significantly correlated to changes on C-MET PET after treatment of rHGG with BEV. - Long PFS could be predicted by T/N ratio at 8 weeks on C-MET and the change in T/N from baseline to 4 weeks follow-up on C-MET and ASL |
| Humbert et al., 201929 | Static F-DOPA Clinical trial and Prospective |
12 | BEV | NA | SUV | Demonstrated the application of F-DOPA PET and MRI in tumor board discussions to implement decisions regarding patient treatment plans - 33.3% cases of suspected tumor recurrence had changed diagnosis and treatment plan with inclusion of F-DOPA PET results |
ASL = arterial spin labeling perfusion imaging; AUC = area under the curve; BEV = bevacizumab; BTV = biological tumor volume; C-MET = [11C]MET; F-DOPA = [18F]FDOPA; F-FET = [18F]FET; IR = irinotecan; LOM = lomustine; MLR = multiple linear regression; MTV = metabolic tumor volumes; NA = not applicable; OCEBM = Oxford Center for Evidence Based Medicine; OS = overall survival; PFS = progression-free survival; PRM = parametric response map; PRS = post-recurrence survival; rHGG = recurrent high-grade glioma; ROC = receiver operating characteristic; rOS = recurrent OS; SUV = standardized uptake value; TAC = time activity curve; TBR = tumor-to-background ratio; TMZ = temozolomide; tOS = total OS; TTP = time to peak; VOI = volume of interest.
aNumber of patients indicates number treated with bevacizumab, not necessarily all patients analyzed in study.