Daniel F. Kelly

BACKGROUND AND OBJECTIVES: The nasoseptal flap (NSF) has been a versatile reconstructive option for extended endonasal skull base surgery, significantly decreasing postoperative cerebrospinal fluid (CSF) leakage rates. One failure mechanism of concern is NSF necrosis. It has been postulated that immediate postoperative MRI flap enhancement can predict flap necrosis. This retrospective study analyzes NSF enhancement to assess for flap viability and CSF leakage. METHODS: Patients from 2012 to 2020 who underwent extended endoscopic endonasal surgery with NSF reconstruction were assessed. Immediate postoperative MRI and delayed 3-month MRI were compared for NSF enhancement. Enhancement was graded as no enhancement, partial, or complete enhancement. Patient demographics, tumor type, intraoperative CSF leak grade, and postoperative CSF leakage were assessed based on flap enhancement patterns. RESULTS: Of 713 patients who underwent endoscopic endonasal surgery, 64 required NSF reconstruction. On the immediate postoperative MRI, 45 patients (70%) had complete flap enhancement, 9 (14%) had partial, and 10 (16%) no enhancement. On the 3-month MRI, 59 patients (92%) had complete flap enhancement and 5 (8%) had partial enhancement. There was significant improvement of flap enhancement between immediate postoperative and 3-month MRI (P = .002). All patients with no initial enhancement had complete enhancement at 3 months. Of those with partial enhancement, 2 remained partial and 7 had complete enhancement at 3 months. Overall, 44 patients (69%) had no change between MRI scans, 17 (27%) improved, and 3 (5%) had decreased enhancement. There was no correlation between intraoperative CSF leak rates and flap enhancement. Four patients had postoperative CSF leaks, 2 having complete immediate enhancement, 1 partial, and 1 without enhancement (P = .85). CONCLUSION: Overall, immediate postoperative MRI NSF enhancement (or lack thereof) did not predict enhancement at the 3-month MRI and did not correlate with postoperative CSF leakage. Hence, one should not rely solely on postoperative flap enhancement to assess the viability of the dural reconstruction.

BACKGROUND AND OBJECTIVES: Postoperative cerebrospinal fluid (CSF) leaks and meningitis are well-known risks of retromastoid craniotomy. Use of abdominal fat grafts, collagen allografts, and rigid or semirigid buttresses have demonstrated efficacy in preventing CSF leaks and meningitis in endoscopic endonasal surgery. This study aims to determine the utility of a similar multilayered reconstruction technique for retromastoid craniotomy. METHODS: We retrospectively reviewed 212 consecutive patients who underwent retromastoid craniotomy for tumor removal or microvascular decompression from 2007 to 2022. Scalp incisions were linear or slightly curved, muscle and facia opening was performed sharply avoiding monopolar cautery; craniotomies had a maximum dimension of 3 cm. A primary water-tight dural closure was rarely achieved favoring collagen sponge overlay often augmented with autologous fat. Clinical factors including pathology, mastoid air cell entry, and reconstruction material were analyzed. Outcomes including postoperative CSF leakage and meningitis were assessed. RESULTS: Of 212 patients (mean age 56 ± 16 years; 60% female; 10% with prior surgery), 148 (70%) had tumor resection and 64 (30%) had microvascular decompression. Mastoid air cells were breached in 67%. Collagen sponge dural overlay was used in 201/212 (95%). A fat graft was placed in 116 (55%) cases: 69% with air cell entry, 27% without air cell entry; 158 (75%) patients had their bone flap replaced, 46 (21%) had titanium mesh cranioplasty, 8 (4%) had no bone flap or titanium mesh. There were no CSF leaks or meningitis. One patient had a lumbar drain placement preoperatively, none postoperatively. Median length of stay was 2 days. CONCLUSION: Retromastoid craniotomy multilayered reconstruction with liberal use of collagen sponge and abdominal fat grafts seems to reliably avoid postoperative CSF leaks and meningitis including in the setting of nonwatertight dural closure and mastoid cell entry. Use of shorter incisions, avoidance of monopolar cautery, and a relatively small craniotomy may contribute to the absence of CSF leaks in this series.

Table S2. Individual listing of additional chordoma tissue samples collected for miR WTA.

Purpose:Chordomas are ultrarare tumors of the axial spine and skull-base without approved systemic therapy. Most chordomas have negative expression of thymidylate synthase (TS), suggesting a potential for responding to the antifolate agent pemetrexed, which inhibits TS and other enzymes involved in nucleotide biosynthesis. We evaluated the therapeutic activity and safety of high-dose pemetrexed in progressive chordoma.Patients and Methods:Adult patients with previously treated, progressive chordoma participated in an open-label, single-institution, single-arm, pilot clinical trial of intravenous pemetrexed 900 mg/m2 every 3 weeks and supportive medications of folic acid, vitamin B12, and dexamethasone. The primary endpoint was objective response rate according to RECIST v1.1. Secondary endpoints included adverse events, progression-free survival (PFS), tumor molecular profiles, and alterations in tissue and blood-based biomarkers.Results:Fifteen patients were enrolled and the median number of doses administered was 15 (range, 4–31). One patient discontinued treatment due to psychosocial issues after four cycles and one contracted COVID-19 after 13 cycles. Of the 14 response-evaluable patients, 2 (14%) achieved a partial response and 10 (71%) demonstrated stable disease. Median PFS was 10.5 months (95% confidence interval: 9 months–undetermined) and 6-month PFS was 67%. Adverse events were expected and relatively mild, with one grade 3 creatinine increased, and one each of grade 3 and 4 lymphopenia. No grade 5 adverse events, unexpected toxicities, or dose-limiting toxicities were observed. Several patients reported clinical improvement in disease-related symptoms.Conclusions:High-dose pemetrexed appears tolerable and shows objective antitumor activity in patients with chordoma. Phase II studies of high-dose pemetrexed are warranted.

Table S6. Representativeness of Study Participants

Table S5. The top 38 cfmiRs that are commonly DE in chordoma vs normal plasma samples when comparing top 100 most detected, top 100 most significant, and top 100 with highest FC.

Supplementary Figure S4. Significant cfmiRs found in plasma and tissues. A. Venn diagram showing the cfmiRs differentially expressed detected in plasma of chordoma patients that were also detected in tissue samples from chordoma patients (normalized counts >30). B. Venn diagram showing the common cfmiRs found in plasma samples considering the top 100 cfmiRs most detected, most significant changing, and most upregulated that were identified for chordoma vs normal healthy donors (NHDs). 38 cfmiRs were commonly identified as the most significantly changing, upregulated, and detected in chordoma patients.

Table S3. Plasma and tissue samples from clinical trial participants used for miR WTA.

Supplementary Figure S2. MRI and CT images of patient P04 with metastatic spinal chordoma in C-spine and lung. C-spine MRI before treatment (A-B) shows C-spine disease (arrows) and 12 months after initiation of high-dose pemetrexed (D-E) shows shrinkage of enhancing tumor on sagittal post contrast T1 sequences (A, D), sagittal T2 (B, E). Chest CT before treatment (C) showed multiple small metastases at baseline axial post contrast sequences that resolved after treatment (F).

Supplementary Figure S1. MR images of patient P05 with locally advanced clival chordoma. Brain MRI before treatment (A-C) shows left cavernous since disease (arrows), 6 months after initiation of high-dose pemetrexed (D-F), and 12 months after treatment initiation (G-J) showing shrinkage of enhancing tumor on axial post contrast T1 sequences (A, D, G), axial FLAIR (B, E, H) and axial T2 (C, F, I).