FLEX January 2024

10976817, 2021, 2, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599820938660 by UNIVERSITY OF MINNESOTA 170 WILSON LIBRARY, Wiley Online Library on [24/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

394

Otolaryngology–Head and Neck Surgery 164(2)

Radiographic STR was defined as any postoperative contrast-enhancing tissue that is correlated with intraopera tive remnant location. Radiographic GTR was defined as an absence of any definitive evidence of residual or recurrent tumors on postresection, surveillance imaging. Statistical Analysis Descriptive statistics were used to summarize categorical variables and continuous variables. Continuous variables were tested for normalcy with the Shapiro-Wilk test, and those found to have a nonnormal distribution were described by median and interquartile range (IQR). Single predictor binary logistic regression models were used to assess poten tial predictors of post-STR tumor growth and the need for postsurgical radiation. Single predictor linear regression was used to evaluate the association of variables with continuous outcomes. Kruskal-Wallis test and Fischer’s exact test evalu ated clinicopathologic variables for association between cohorts. Single predictor Cox proportional hazards regression models were constructed to assess variables for association with time to salvage therapy. Proportionality was assessed for all covariates by the Kolmogorov-type supremum test with 1000 simulations and by graphing Martingale score residuals versus time. Salvage-free survival was analyzed by the Kaplan–Meier method. Multivariable analysis was not per formed because of the limited number of disease events. All confidence intervals were reported at 95%, and all P values were reported as two-sided, with an alpha level of .05. All statistical analyses were performed using SAS ver sion 9.4 (SAS Institute Inc., Cary, NC), and plots were gen erated using Prism version 7 (GraphPad Software, La Jolla, CA). Results Between July 2006 and March 2019, 41 patients (median age 47, 76% female) were retrospectively identified ( Table 1 ). Seven patients had previously undergone treatment at other institutions at the time of presentation, including 4 (10%) with prior microsurgical resection and 3 (7%) having received radiation. Four patients (10 %) had a family history of PGL, and 10 (24%) patients had another synchronous PGL. Pulsatile tinnitus (73%) and hearing loss (80%) were the most common presenting symptoms. Most patients pre sented with advanced-stage disease as categorized by both Glasscock-Jackson (76% with stage III and IV; Table 1 ). Surgical Approach and EOR Most surgical patients (78%) underwent STR ( Table 2 ). Extended STR (type 1) was employed in 19 (59%) patients in the STR cohort. All patients in the GTR cohort had a combined neurotologic and cervical approach, compared with 16 (50%) patients in the STR cohort who underwent isolated neurotologic procedure. The most common neuroto logic procedure was IFTA, with fallopian bridge technique (39%), followed by tympanomastoidectomy (26%). Twenty one patients in the STR cohort had both pre- and postoperative imaging of sufficient quality to generate actuarial volumetric

Table1. Patient Demographics and Disease Characteristics.

Median/ incidence

Variable

IQRor%

Age (years)

47 31

35-59

Gender (female) Prior surgery Prior radiation

76% 10%

4 3 4 3 4 4 3 2

7%

Family history of PGL

10% 24%

Other PGL

10

Contralateral jugular Ipsilateral carotid body Contralateral carotid body

7%

10% 10%

Ipsilateral vagal Other location

7% 5%

Most common symptoms at presentation Pulsatile tinnitus

30 33 11

73% 80% 26% 22% 32%

Hearing loss Disequilibrium

Otalgia

9

Aural fullness

13

Glasscock-Jackson staging I

2 8

5%

II

20% 27% 49%

III IV

11 20

Laterality Right

22 19 35

54% 46%

Left

Postoperative follow-up (months)

10-87

Abbreviations: IQR, interquartile range; PGL, paraganglioma.

data. The median EOR based on volumetric analysis was 36% (0%-96%). On average, EOR was smallest for patients following type 3 STR, with a median EOR of 17% (IQR 4%-29%). In contrast, the median EOR following type 1 STR was 48% (IQR 28%-68%) and 64% (IQR 55%-72%) after type 2 STR. However, the type of STR was not statisti cally correlated with EOR ( p = 0.06). Table 3 expands on the clinical and tumor parameters between extended STR (type 1) and Type 2/3. There was no difference in terms of initial tumor stage (Glasscock-Jackson) or preoperative tumor volume between cohorts ( Table 3 ). Tumor Control The median postoperative follow-up was 35 months (IQR, 10-87 months). Fifteen patients in the STR (46.9%) had tumor growth following initial surgery (type 1, 68%; type 2, 20%; type 3, 12.5%). One patient had recurrence following GTR (11.1%). Binary logistic regression demonstrated STR (type 1) to be associated with risk of tumor regrowth com pared with type 2 and 3 (odds ratio [OR] 11.9, 95% confi dence interval [CI] 1.99-71, p = 0.007). There were no other significant clinical, disease or treatment-related predictors for post-STR tumor growth ( Table4 ).