Improved Detection of Postoperative Residual Meningioma with [⁶⁸Ga]Ga-DOTA-TOC PET Imaging Using a High-resolution Research Tomograph PET Scanner

An important aspect in meningioma management following primary surgery is predicting the risk of recurrence. In older literature, the degree of surgical resection assessed by the Simpson grading scale has been an important prognostic factor for recurrence rate ranging between 9% and 19% at 10 years after gross-total resection (GTR; Simpson grades 1–3) and up to 44% for subtotal resection (Simpson grade 4), even when the tumor is histologically benign [World Health Organization (WHO) grade I; ref. 1]. Over the recent years, however, the Simpson grading scale has been reported to have a poor reproducibility, and it is debatable whether the recurrence rate is directly linked to the grading scale (1–3). Tumor location has also been considered a predictor of recurrence, negatively impacting the extent of resection and thus the recurrence-free survival, for example, meningiomas located at the skull base with their proximity to the critical neurovascular structures, in the orbits, or near venous sinuses (4, 5). Finally, transosseous extension of meningioma is another important risk factor for tumor recurrence (6, 7), as it frequently impedes GTR because of the insufficient demarcation of osseous involvement (8, 9). According to the European Association of Neuro-Oncology (EANO) guidelines (10), the current clinical standard for the assessment of tumor resection is contrast-enhanced MRI, which is routinely performed at 3-month follow-up, and hereafter annual MRI controls until 5 years after treatment, then every 2 years. However, conventional MRI is of suboptimal sensitivity and specificity when distinguishing scar tissue from meningioma residue(s) because of nonspecific reactive signal abnormalities in both cases. Other challenges in MRI are identification of residues because of osseous or muscular infiltration, locations in skin or cavities, or near or invading vascular structures (e.g., dural sinuses), as well as residues obscured by MRI susceptibility artefacts due to hemorrhage, or metallic implants (e.g., craniotomy fixation and titanium net). A precise knowledge of possible meningioma residues is imperative, and may have a significant impact in determining subsequent patient management.

PET with the somatostatin receptor analog DOTA-D-Phe¹-Tyr³-octreotide labeled with Gallium-68 ([⁶⁸Ga]Ga-DOTA-TOC) offers an additional tool for imaging meningiomas. This molecular imaging method is based on the characteristic of meningioma cells expressing high levels of somatostatin receptors especially subtype 2 (SSTR2), to which [⁶⁸Ga]Ga-DOTA-TOC binds to with very high affinity, providing an excellent tumor-to-background ratio regardless of artefacts caused by metallic implants (10–13). In a histopathologic study, Rachinger and colleagues (14) investigated a similar [⁶⁸Ga]Ga-DOTA–conjugated peptide, [⁶⁸Ga]Ga-DOTA-D-Phe¹-Tyr³-octreotate ([⁶⁸Ga]Ga-DOTA-TATE), and reported a significant association between tracer uptake using maximum standardized uptake value (SUV_max) and SSTR2 assessed IHC from 115 tissue samples (81 tumors, 34 tumor free) in 21 patients with meningioma. Recently, we examined the correlative relationship using a range of simplified and fully quantitative [⁶⁸Ga]Ga-DOTA-TOC PET metrics and SSTR2 expression using both IHC and quantitative PCR in a cohort of 15 patients, and found significant strong positive correlations as well (15). Numerous studies have already reported increased sensitivity of tumor detection leading to significant modifications of the planned target volume delineation in connection with radiotherapy planning after [⁶⁸Ga]Ga-DOTA-TOC or [⁶⁸Ga]Ga-DOTA-TATE PET (16–22). In addition, an improved detection of osseous involvement has been reported with [⁶⁸Ga]Ga-DOTA-TATE in preoperative and postoperative clinical timepoints in contrast to MRI (23). However, information regarding its use for postoperative assessment solely is limited (24), and thus, the evaluation of the usefulness of this imaging modality in the postoperative setting seems reasonable.

In this single-center study, we aimed to investigate the diagnostic accuracy of [⁶⁸Ga]Ga-DOTA-TOC PET imaging as a supplement to 3-month postoperative MRI reporting GTR. For the study purpose, we employed the use of a brain-dedicated high-resolution research tomograph (HRRT) PET scanner with an image resolution two to three times superior than compared with other PET or PET/CT scanners used in daily practice, making it ideal for delineation of even very small tumor residues (20).

The study protocol was approved by the Regional Ethics Committee in Copenhagen, Denmark (H-15006091), and conducted in accordance to the Declaration of Helsinki II. All patients had given written informed consent before the study enrollment. Between November 2017 and May 2019, a group of 37 consecutive patients with meningioma [29 females, 8 males; median age 66 years (range, 28–78 years)] with gross-totally resected meningioma according to 3-month postoperative MRI studies were prospectively enrolled, and referred to [⁶⁸Ga]Ga-DOTA-TOC PET. Clinical information up to the time of [⁶⁸Ga]Ga-DOTA-TOC PET was retrospectively obtained using patient medical records, radiological data, and histopathologic specimens. No patients had undergone preoperative embolization. Three patients (8.1%) were treated for a recurrent meningioma. Tumor resections were performed using the neuronavigation system based on MRI data, and the resection extent was classified using the Simpson grading scale (1). When not explicitly stated, Simpson grades were assigned on the basis of the details of the operative report. However, while no patient was excluded because of a missing Simpson grade, patients with radiological evidence of residual tumor were excluded from participation in the study regardless of the Simpson grade. Simpson grade 1 resection was achieved in 19 patients (51.4%), Simpson grade 2 in 6 patients (16.2%), Simpson grade 3 in 8 patients (21.6%), and Simpson grade 4 in 3 patients (8.1%), while in the last patient (2.7%), the Simpson grade could not be determined. Histopathology revealed WHO grade I meningioma in 28 surgical specimens (75.7%), WHO grade II meningioma in 8 specimens (21.6%), and WHO grade III meningioma in 1 specimen (2.7%). The median time between 3-month postoperative MRI and [⁶⁸Ga]Ga-DOTA-TOC PET was 6 weeks (range, 1–16 weeks). The case with the 16-week scan interval was included, as no progression or obvious tumor residue was observed on follow-up MRI indicating that tumor growth per se was unlikely to account for PET detection. Three patients (8.1%; 2 WHO grade II and 1 WHO grade III meningioma) subsequently underwent adjuvant radiotherapy. Demographic factors, meningioma location and size prior to surgery, existence of a dural tail, Simpson grading scale, and histopathology including subtypes and grades are presented in Table 1.

A dedicated brain HRRT PET scanner (CTI/Siemens) was used for the purpose (25, 26). The axial field of view was of 25 cm and a near isotropic resolution of 1.4 mm. The synthesis of [⁶⁸Ga]Ga-DOTA-TOC is described previously in detail (27). [⁶⁸Ga]Ga-DOTA-TOC was injected with a median dose of 102 MBq (range, 82–111 MBq) as bolus intravenously followed by a tracer uptake phase of approximately 40 minutes. The patient was then positioned in the PET scanner with the head rested in a foam-cushioned headrest and a head strap over the forehead to minimize head movements. Initially, a 6-minute transmission scan with a rotating ¹³⁷Cs single-photon point source was performed for attenuation correction, and subsequently, a single 20-minute PET frame was recorded at 45-minute postinjection. The images were reconstructed using a three-dimensional (3D)-ordered subset expectation maximization algorithm with correction for the measured point spread function (28). Each image consisted of 207 image planes in a 256 × 256 matrix with an isotropic voxel of 1.22 × 1.22 × 1.22 mm³. All images were corrected for randoms, scatter, attenuation, decay, and dead time, and filtered with a 3D Gaussian 2-mm filter (29). The resultant image resolution, when the high [⁶⁸Ga]Ga positron energy (mean 0.84 MeV) was accounted for, was estimated to approximately 3.8-mm full width at half maximum (30).

The preoperative and postoperative T1-weighted MRI and PET data were fused using medical imaging software, MIRADA DBx (2016 Mirada Medical Ltd.). Volumes of PET-enhanced meningioma residue(s) were assessed using a standardized volume of interest (VOI) method, where a VOI was defined enclosing all enhanced lesions with the highest [⁶⁸Ga]Ga-DOTA-TOC accumulation within or adjacent to the surgical cavity, as described later. As no established threshold based on biopsy-controlled confirmation exists for meningioma tissue following [⁶⁸Ga]Ga-DOTA-TOC and the HRRT PET scanner, the threshold established for [⁶⁸Ga]Ga-DOTA-TATE measured as SUV of ≥2.3 was used as a supporting criterion for separating meningioma from nontumoral tissue (14). In addition, the volume of residual meningioma in cm³ was recorded.

All MRI scans were performed on a 1.5 or 3 T MRI scanner with voxel dimensions of ≤1.5 mm at 3 months postoperatively. The scanning protocol included axial T1- and T2-weighted images, and axial, coronal, and sagittal T1-weighted images after intravenous administration of gadolinium-based contrast agent.

The primary endpoint was local meningioma residue detected on [⁶⁸Ga]Ga-DOTA-TOC PET up to 2 cm from the surgical cavity (Table 2). However, distant disease identified on [⁶⁸Ga]Ga-DOTA-TOC PET extending beyond 2 cm from the surgical cavity that was not reported on 3-month postoperative MRI studies, was recorded as well. The Simpson grade was cross-checked against the findings of 3-month postoperative MRI and [⁶⁸Ga]Ga-DOTA-TOC PET scans. The secondary endpoint was verification of these suspicious meningioma residues suggested by [⁶⁸Ga]Ga-DOTA-TOC PET, which was done by either direct histology, or disease recurrence/progression defined as (new) lesions with a ≥3-mm change in the size of the unidimensional maximal diameter on subsequent T1-MRI, compared with 3-month postoperative MRI, according to the latest published endpoint trial criteria from Response Assessment of Neuro-Oncology (RANO) workgroup (31). More importantly, disease progression was determined only if detected at the exact location of the [⁶⁸Ga]Ga-DOTA-TOC-avidity on PET scan. In cases with uncertainty that could not be solved with subsequent MRI scans, a follow-up [⁶⁸Ga]Ga-DOTA-TOC PET was performed at least 6 months following the initial PET, where volume changes were evaluated using identical SUV thresholds, as described above.

Data analysis was performed with the statistical software package IBM SPSS (version 26.0, IBM Corp.). Continuous data were described with median with range. Differences between the results of MRI and PET studies were analyzed using nonparametric Fischer exact or Mann–Whitney U tests, whenever appropriate. The tests were two sided, and the level of significance was set at less than 0.05. The diagnostic accuracy of [⁶⁸Ga]Ga-DOTA-TOC PET in detecting meningioma residues was assessed on a per patient basis using ROC curve analysis. The corresponding sensitivity, specificity, and positive and negative predictive values (PPV and NPV) with 95% confidence intervals (CI) were determined using 999 bootstrapped replications of the data. In addition, the AUC was provided. Because of a relatively short follow-up time, possible factors including [⁶⁸Ga]Ga-DOTA-TOC-avidity related with subsequent disease recurrence using Cox regression analysis were not investigated.

Sixteen postoperative MRI scans showed nonspecific contrast-enhancing lesions near the surgical cavity (Fig. 1). In contrast, a total of 23 [⁶⁸Ga]Ga-DOTA-TOC PET scans (62.2%) showed a solitary or multiple intense [⁶⁸Ga]Ga-DOTA-TOC-avid lesions (n = 29). Twenty-four lesions (82.8%) were suspected to be local tumor residues (Fig. 2), and five (17.2%) were interpreted as additional lesions or distant disease (Fig. 3B), where one was classified as a possibly iatrogenic implantation with a lesion located in the subcutaneous skin at the site of the surgical cavity (Table 2). A high physiologic uptake of [⁶⁸Ga]Ga-DOTA-TOC was found in the pituitary glands of all patients, and did not significantly hamper the PET-based demarcation of adjacent lesions in any case. Transosseous expansion was detected in five cases (13.5%). Six of 37 meningiomas (16.2%) had a dural tail sign preoperatively, of which only one demonstrated an active [⁶⁸Ga]Ga-DOTA-TOC-avid residue. The median SUV_max and SUV_mean were 7.7 (range, 3.6–40.6) and 4.1 (range, 2.3–10.9), and the median PET-based volume was 0.2 cm³ (range, 0.1–6.3 cm³), where 17 lesions (73.9%) were less than 1.0 cm³. A significant association was found between [⁶⁸Ga]Ga-DOTA-TOC-avid lesions and tumor locations (convexity vs. skull base; P = 0.039), with [⁶⁸Ga]Ga-DOTA-TOC-avidity predominately observed at the skull base region (Tables 1 and 2).

As mentioned above, 16 of 37 postoperative MRI scans (43.2%) were described with nonspecific contrast-enhancing areas near the surgical cavity, where six of corresponding operative reports were assigned Simpson grade 1 (Table 3). The remaining 10 MRI scans were attributed to Simpson grade 2 (n = 3), Simpson grade 3 (n = 4), and Simpson grade 4 (n = 3). One MRI scan (2.7%) failed to detect any contrast enhancement in a patient with Simpson grade 4 following the resection of a parafalcine/parasagittal meningioma. Reactive dural enhancement and gliosis was reported in 27 and 14 MRI scans (73% and 37.8%), respectively. The Simpson grade was neither associated with nonspecific contrast enhancement (P = 0.375), reactive dural enhancement (P = 0.847), nor gliosis (P = 0.129).

Of 22 PET studies with [⁶⁸Ga]Ga-DOTA-TOC-avid lesions when excluding the case with the iatrogenic lesion, 10 corresponding operative reports assigned Simpson grade 1, and three PET scans corresponded to Simpson grade 2, five to Simpson grade 3, and four to Simpson grade 4 (Table 3). No association was found between [⁶⁸Ga]Ga-DOTA-TOC-avidity and Simpson grade (P = 0.488).

Concordant negative findings, for example, GTR on MRI and no tracer uptake on PET scans, were recorded in 14 patients (37.8%; Fig. 1). In contrast, there was a significantly higher proportion of [⁶⁸Ga]Ga-DOTA-TOC-avid lesions in the remaining patients with nonspecific findings on MRI (81.3%), compared with those without (47.6%; P = 0.048). No differences were found in tracer uptake or PET-based volume (median SUV_max, 7.7 vs. 7.5, P = 0.821; median SUV_mean, 4.1 vs. 4.1, P = 0.539; and median VOI, 0.20 vs. 0.25 cm³, P = 0.483) between patients with 3-month postoperative MRI studies with and without nonspecific contrast enhancement, respectively. [⁶⁸Ga]Ga-DOTA-TOC-avid lesions adjacent to the superior sagittal sinus or the edge of craniotomy (especially the anterior part) on the convexity (Fig. 2A and B), as well as the lesions located in the anterior cranial fossa of the skull base particularly, were difficult to detect on 3-month postoperative MRI (Fig. 3A). For instance, five [⁶⁸Ga]Ga-DOTA-TOC-avid lesions adjacent to the superior sagittal sinus and four lesions related to the anterior edge of a craniotomy could not be detected (Fig. 3B; Table 2).

The [⁶⁸Ga]Ga-DOTA-TOC-avidity solely did not result in a prompt change in patient management in any case. The median length of follow-up time was 17 months (range, 9–35 months) with the last MRI follow-up on November 13, 2020. Overall, a total of 10 patients (6 WHO grade I and 4 WHO grade II) experienced subsequent progression (Figs. 2 and 3), whereas subsequent disease progression was observed predominantly in patients with [⁶⁸Ga]Ga-DOTA-TOC-avid areas (n = 8), compared with 2 patients without tracer uptake (P < 0.0001; Fig. 2D). Upon assessing WHO grade I meningiomas (n = 28) only, a total of 14 PET scans showed [⁶⁸Ga]Ga-DOTA-TOC-avid lesions, where five of them demonstrated disease progression, compared with one without preceding PET-avidity (P < 0.0001). No significant differences were seen with regard to PET-based volume or tracer uptake between confirmed stable and progressive residues (median SUV_max, 6.2 vs. 9.6, P = 0.142; median SUV_mean, 3.8 vs. 4.3, P = 0.681; median VOI, 0.25 vs. 0.30 cm³, P = 0.743). Of patients that had undergone adjuvant radiotherapy (8.1%), only one showed [⁶⁸Ga]Ga-DOTA-TOC-avidity, and none of them experienced recurrence. However, the latter must be interpreted with caution, as this might be related to treatment response.

Eighteen of 23 PET scans could be verified using the references available (Fig. 1). In total, 8 patients (44.4%) demonstrated significant progression of their [⁶⁸Ga]Ga-DOTA-TOC-avid lesions during follow-up time as mentioned above, where 2 of them (25%) underwent re-resection with subsequent histology confirming a meningioma tissue. In another patient (2.7%), a meningioma residue was discovered, located in the subcutaneous skin at the site of surgical cavity at the time of radiotherapy planning that was completely missed on 3-month postoperative MRI, but later verified histologically. In the remaining patients (50%) after obtaining subsequent informed consent, a follow-up [⁶⁸Ga]Ga-DOTA-TOC PET was performed at a median of 22 months (range, 9–35 months) after the initial PET scan that confirmed tracer binding at identical locations in all cases, supporting “tumor residue” rather than just “nonspecific uptake” (Figs. 2 and 3). The active tracer volume was increased in six cases, with a median of 0.3 cm³ (range, 0.2–0.6 cm³).

Upon excluding patients with [⁶⁸Ga]Ga-DOTA-TOC uptake without subsequent confirmation or follow-up (n = 6; Fig. 1), the ROC analysis yielded a diagnostic accuracy of 90% [(95% CI, 74–98), sensitivity 90% (95% CI, 67–99), specificity 92% (95% CI, 62–100), PPV 94% (95% CI, 72–99), NPV 85% (95% CI, 59–95), and AUC of 0.906 (95% CI, 0.783–1.000); P < 0.0001] for [⁶⁸Ga]Ga-DOTA-TOC PET in detecting small unrecognized residual meningioma.

In this study, we evaluated the diagnostic accuracy of [⁶⁸Ga]Ga-DOTA-TOC PET for postoperative assessment of residual meningioma tissue. The study confirmed our hypothesis that the increased sensitivity of [⁶⁸Ga]Ga-DOTA-TOC PET would increase the detection rate of tumor tissue, as found in 62.2% of patients. Particularly those with nonspecific findings on MRI were more likely to obscure [⁶⁸Ga]Ga-DOTA-TOC-avid lesions (Figs. 2 and 3). In addition, one [⁶⁸Ga]Ga-DOTA-TOC-avid lesion located in the subcutaneous skin was discovered only in connection with [⁶⁸Ga]Ga-DOTA-TOC PET at the time of radiotherapy planning, and was confirmed as a meningioma residue by histology. Furthermore, [⁶⁸Ga]Ga-DOTA-TOC PET revealed transosseous extension in 5 patients (13.5%), where 2 of them progressed subsequently. Finally, the initial tracer uptake was post hoc deemed reliable, as the findings were confirmed by a second [⁶⁸Ga]Ga-DOTA-TOC PET scan after a median follow-up period of 22 months, limiting the potential risk of nonspecific uptake following surgery (Fig. 3). Upon excluding patients where verification of PET findings and follow-up was not possible (Fig. 1), the estimate of the diagnostic accuracy of [⁶⁸Ga]Ga-DOTA-TOC PET was 90% with a sensitivity and specificity of 90% and 92%, respectively. Thus, [⁶⁸Ga]Ga-DOTA-TOC PET could be a useful imaging modality as a supplement to MRI in planning of individual patient management.

Contrast-enhanced MRI is currently the gold standard imaging modality with meticulous respect for details in both diagnostic evaluation and treatment planning (10). However, the diagnostic accuracy of MRI is limited in terms of assessing GTR, especially at complex anatomic locations such as in the skull base and parasagittal regions where invasion of bone or dural sinus is frequently present (12, 13). In addition, it is well known that subtotally resected meningiomas are likely to recur regardless of their histologic appearance. In this study, a low concordance (37.8%) was found between postoperative MRI and [⁶⁸Ga]Ga-DOTA-TOC PET scans with regard to tumor resection completeness (Fig. 1). [⁶⁸Ga]Ga-DOTA-TOC-avidity was predominantly seen at the skull base area, followed by within or near the superior sagittal sinus, and just below the edge of craniotomy (Table 2), which was not always radiologically detectable. For instance, in 1 patient with Simpson grade 4 following the resection of parafalcine/parasagittal meningioma (2.7%), the 3-month postoperative MRI scan was unable to demonstrate any pathologic contrast enhancement (Fig. 2A). Similarly, Afshar-Oromieh and colleagues (32) compared contrast-enhanced MRI and [⁶⁸Ga]Ga-DOTA-TOC PET/CT prior to radiotherapy, and reported that all meningiomas (n = 190) could be detected by PET/CT, while only 171 meningiomas were detected by contrast-enhanced MRI (90%), because small tumors, in particular those adjacent to falx cerebri, location in the skull base, or obscured by imaging artefacts or calcification were missed by MRI, as similarly illustrated in our study (Fig. 3B).

Of patients with [⁶⁸Ga]Ga-DOTA-TOC-avid lesions, about one third experienced disease progression. One may speculate whether the tracer-avidity is a surrogate marker of tumor aggressiveness. In a recent PET study with similar DOTA-analog, [⁶⁸Ga]Ga-DOTA-TATE, Sommerauer and colleagues (33) reported a strong association between tracer uptake using SUV_max and disease progression assessed as tumor growth rate (r = 0.757, P < 0.001) that was irrespective of WHO grades. However, this association may merely reflect referral bias as the sample had an overrepresentation of patients with aggressive features such as WHO grade II and III (∼70%), multifocality, and transosseous growth, and thus cannot be used to support causality between SSTR2 expression and growth. In fact, the prediction would be that the SSTR2 expression infers growth inhibition (34). Moreover, recurrent meningiomas with known aggressive phenotypes can arise from even minor remnants below the PET detection level, as reported in our study. In 2 patients (5.4%), [⁶⁸Ga]Ga-DOTA-TOC PET scans were classified as falsely negative, where one of the patients had undergone surgery (Simpson grade 1) for a recurrent WHO grade II meningioma, and experienced disease recurrence 17 months later (Fig. 2D). This particular patient had an intensely avid [⁶⁸Ga]Ga-DOTA-TOC uptake preoperatively (image not shown), indicating that the remnant on postoperative PET scan must have been rather small. Thus, in cases with recurrent meningiomas with known aggressive phenotypes, the patients must be followed with closer monitoring regardless of postoperative imaging findings. Another concern could be that tumor growth occurred in other locations in the resection cavity than the tracer active area. However, in all cases with MRI-based progression, this was correlated closely with the initial PET findings, as defined in the Materials and Methods section.

Whether there are certain biological alternations of meningiomas related to their location that makes them susceptible to recurrence is not known to date. Especially the close proximately to the falx cerebri and/or superior sagittal sinus has been reported to preclude radical surgery leading to higher recurrence rates (35, 36), as illustrated in Fig. 2A, although the results are not always consistent throughout the literature (5, 37). We found that tumor locations such as skull base and/or parasagittal regions were a significant risk factor for incomplete resections, where a Simpson grade 1 in particular is already difficult to achieve.

The potential of [⁶⁸Ga]Ga-DOTA-analog in the setting of postoperative meningioma already has been investigated in a few studies (24, 38). One study evaluated [⁶⁸Ga]Ga-DOTA-TATE using a PET/MRI scanner in 17 patients that had undergone subtotal resection with and without adjuvant radiotherapy (24). The authors reported a significant differentiation between meningioma and postoperative changes based on SUV_max normalized to the blood radioactivity obtained from the superior sagittal sinus (16.6 vs. 1.6, P < 0.0001). Similarly, [⁶⁸Ga]Ga-DOTA-TATE PET revealed additional small meningiomas that were not prospectively identified on conventional MRI in 4 patients. In contrast, only patients with gross-total resection on postoperative MRI were included in this study. Using a threshold of at least 2.3, [⁶⁸Ga]Ga-DOTA-TOC PET exhibited a rather high tracer uptake in all suspected lesions despite a possible risk of underestimation of SUVs in small tumor volumes owing to the partial volume effects (PVE). Thus, the findings in this study are reliable despite lack of histopathologic confirmation in all cases and a relative short follow-up time. The threshold of 2.3 was implemented from a previous biopsy-controlled study to support the presented results, as it has been reported to be optimal for [⁶⁸Ga]Ga-DOTA-TATE for discriminating between meningioma and nontumoral tissue (14). In a subsequent study conducted from the same institution, Ueberschaer and colleagues (38) evaluated the intraoperative estimation of the extent of tumor resection using the Simpson grade with reference to postoperative MRI and [⁶⁸Ga]Ga-DOTA-TATE PET/CT in 49 patients with a total of 52 WHO grade I meningioma, where a total of 37 were judged as complete resections intraoperatively (Simpson grades 1–2). Likewise, the authors reported the superiority of [⁶⁸Ga]Ga-DOTA-TATE PET in terms of tumor resection, with tumor residuals predominantly found in the convexities and falx, followed by the skull base region. The object of this study was to assess the value of Simpson grade, and thus 21% of patients had a partial resection (Simpson grade 4), while our primary target was to assess the diagnostic value of a GTR reading (Simpson grades 1–3) on MRI according to the latest EANO recommendations. In Simpson grade I, they reported 15/37 (40.5%) PET scans with [⁶⁸Ga]Ga-DOTA-TATE-avidity, which is consistent with the proportion in our study of 14/28 (50%), when the higher resolution of our PET scanner is taken into consideration. A subgroup of their patients (10/17; 58.8%) resembling our cohort with GTR on postoperative MRI identified similar residual volumes on postoperative PET scans [median BTV: 0.6 mL (0.1–5.3 mL) vs. 0.2 mL (0.1–6.3 mL)]. Furthermore, the diagnostic accuracy of [⁶⁸Ga]Ga-DOTA-TATE PET in identifying meningioma residuals was not pursued in the German study. Finally, MRI and PET scans were performed at late timepoints postoperatively in our study, where diagnostic accuracy is less likely to be influenced by postoperative changes (median 3 and 4.5 months vs. 4 and 25 days). Overall, our findings confirm and extend subresults from Ueberschaer and colleagues (38), but addresses more directed the criteria in the present MRI-focused EANO guidelines.

Although it is recommended that the intraoperative assessment of the extent of resection should be confirmed by a 3-month postoperative MRI (10), our data indicated that the surgical assessment, including incomplete resection (Simpson grade 4) as reported in this study, was more accurately assessed with [⁶⁸Ga]Ga-DOTA-TOC PET than postoperative MRI, which may help in planning subsequent individualized imaging surveillance and/or adjuvant therapy especially in cases of progressive and high-grade meningioma, as well as in stratification into clinical trials to reduce misclassification bias (10, 39). It has recently been shown that the addition of PET imaging for target volume definition in radiotherapy planning led to an improved delineation of tumor tissue that translated into a significantly enhanced local control in WHO grade I meningiomas compared with treatment planning with MRI and CT alone (40). In addition, retrospective studies have shown prolonged survival from adjuvant stereotactic radiosurgery when delivered promptly after resection rather than at regrowth, emphasizing the importance of a sensitive and specific noninvasive imaging technique in patient selection, radiotherapy planning, and subsequent imaging surveillance (10, 12, 41). Furthermore, the routine clinical use of [⁶⁸Ga]Ga-DOTA-TOC PET may be particular important in patients with meningioma with GTR on MRI, where the molecular features, for example, TERT mutation or a progressive phenotype based on next-generation sequencing or clinical behavior would support early radiotherapy (35, 42). Finally, the high prevalence of active tumor residue has implications for the inclusion of patients with allegedly reported GTR on MRI into randomized controlled trials (43), which may in turn lead to erroneous conclusions regarding treatment efficacy, e.g., if tumor residues are either missed or not located in the radiation treatment field.

There are some limitations inherent to this study that need to be considered. First, we did not put much emphasis on PET metrics and calculations of thresholds because of the small volume of [⁶⁸Ga]Ga-DOTA-TOC-avid lesions, as the influence of PVE would lead to inaccurate quantitative estimations, and the uptake thus was merely used as a detection marker. Second, the Simpson grade was not available in all cases and determined on the basis of the surgical reports. Although the majority of procedures were performed by the coauthors (K. Fugleholm and M. Ziebell), the Simpson grading scale is not used consistently by all surgeons at our institution owing to its controversy and the numerous revolutions in imaging, molecular biology, and stereotactic radiotherapy since its introduction in 1957 (1–3, 44). Third, the results were not confirmed directly by biopsy in all cases, but relied on identification of progression on subsequent MRI scans of tumors with variable growth rates (with histologic verification in a subgroup) and PET scans with confirmation of [⁶⁸Ga]Ga-DOTA-TOC-avid lesions, as the former would neither be medically indicated nor ethical. However, the latter may present a risk for incorporation bias given the absence of other accepted reference standard used such as MRI follow-up and/or histology and thus a subsequent risk for the overestimation of diagnostic accuracy. That being said the use of MRI is contingent on meningioma growth, and is inherently slow and with low sensitivity. The main concern with nonspecific [⁶⁸Ga]Ga-DOTA-TOC uptake would be postoperative reactive changes that, however, are unlikely to persist on follow-up PET scans performed more than median 12 months after resection. We thus regard this source of bias to have little impact on the results. A longer follow-up could provide additional prognostic information about tumor recurrence/progression. Finally, a single case with an iatrogenic lesion located in skin, as well as two cases with subsequent progression were all confirmed histologically, which underlines the utility of [⁶⁸Ga]Ga-DOTA-TOC PET in the postoperative setting. The strength was the use of a PET scanner with a very high resolution and confirmation of [⁶⁸Ga]Ga-DOTA-TOC-avid lesions in the majority of cases. Moreover, it should be highlighted that the patients included in this study differed from the remaining PET literature with [⁶⁸Ga]Ga-DOTA–conjugated peptides in meningioma. Herein, the patients were identified as tumor free on 3-month postoperative MRI scans, where a PET scan would usually not be indicated, while the literature consists of patients with MRI verified unresected or remnant meningiomas referred for radiotherapy planning (16–18, 20, 40, 45, 46).

Overall, we confirm that [⁶⁸Ga]Ga-DOTA-TOC PET is a highly effective method for detection of meningioma residues and their distinction from other types of tissue such as postoperative scarring with a significant high diagnostic accuracy (90%). Thus, we recommend the extended use of [⁶⁸Ga]Ga-DOTA-TOC PET in the postoperative setting with patients that have undergone surgical resection for meningioma as a supplement to the 3-month postoperative MRI. It could help to confirm a diagnosis of meningioma residue and thereby influence the subsequent treatment planning and follow-up, as well as the eligibility of patients with meningioma considered for randomized controlled trials. The superior sensitivity needs to be proven longitudinally with serial follow-up MRI studies.

No disclosures were reported by the authors.

A. Bashir: Conceptualization, resources, data curation, software, formal analysis, funding acquisition, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. V.A. Larsen: Software, formal analysis, visualization, writing–review and editing. M. Ziebell: Conceptualization, methodology, writing–review and editing. K. Fugleholm: Conceptualization, methodology, writing–review and editing. I. Law: Conceptualization, resources, data curation, software, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–review and editing.

The authors would like to thank the technical staff, Elin Maria Lindell, Anna Elisabeth Ljunggren, and Christiana Schulze, and radiochemists at the Department of Clinical Physiology, Nuclear Medicine and PET, for help with PET scanning and patient care. We would also like to thank The John and Birthe Meyer Foundation, who generously donated the HRRT PET scanner and the cyclotron to Copenhagen University Hospital, Rigshospitalet. Finally, a special thanks to the statistician, Lasse Anderberg, for aid with the ROC analyses. The work was supported by a grant from the Danish Cancer Society Research Center (R146-A9508-16-S2), where the last author, I. Law, was directly supported by the grant.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.