{"id":"c26a5633-a8a9-48a3-a59a-bfbcdab19d08","slug":"anteroposterior-suprasellar-diameter-as-a-predictor-of-gross-total-resection-in-pituitary-adenomas-with-suprasellar-extension-a-retrospective-cohort-study","title":"Anteroposterior suprasellar diameter as a predictor of gross total resection in pituitary adenomas with suprasellar extension: A retrospective cohort study","authors":["Juan Antonio García Martínez","Ángel Adrián Soto Polendo","José Antonio Candelas Rangel","Gervith Reyes Soto","Carlos Castillo Rangel","Tshiunza Cherubin","Ismael Antonio Peralta Baez","Manuel De Jesus Encarnacion Ramírez"],"abstract":"Background: Pituitary adenomas (PAs) with suprasellar extension pose significant surgical challenges due to restricted access and a heightened risk of subtotal resection (STR). To test this hypothesis, we evaluated suprasellar morphometrics focusing on anteroposterior suprasellar diameter(APSS) as predictors of gross total resection (GTR) in PAs with suprasellar extension, while accounting for parasellar invasion using established cavernous sinus scales (Knosp and Zurich pituitary score [ZPS]). Methods: A retrospective cohort of 48 consecutive patients who underwent endoscopic endonasal surgery for PAs with suprasellar extension between July 2021 and June 2024 was analyzed. Demographic, radiologic, and intraoperative variables were collected. The extent of resection was determined by early postoperative magnetic resonance imaging. Logistic regression and receiver operating characteristic (ROC) curve analysis were used to identify predictors of GTR. Results: The mean patient age was 52.2 ± 11.6 years, with a male-to-female ratio of 1.1:1. Most tumors were nonfunctioning (81.3%). GTR was achieved in 70.8% of cases. Postoperative complications occurred in 27.1%, including cerebrospinal fluid leaks in 16.7%, one postoperative hematoma (2.1%), and two deaths (4.2%). Neither Knosp nor ZPS classifications showed significant correlation with GTR (P = 0.055 and P = 0.258, respectively). Logistic regression identified APSS as an independent predictor of resectability (P = 0.023). ROC analysis established an optimal APSS cutoff of 25.7 mm, with 78% sensitivity and 88% specificity for predicting STR. Knosp and ZPS are designed to quantify cavernous sinus invasion; in our suprasellar-extension cohort, their association with GTR did not reach statistical significance, whereas APSS capturing sagittal suprasellar depth beyond the endonasal working angle remained an independent predictor of incomplete resection. Conclusion: The APSS is an independent and reliable morphometric predictor of surgical outcome in PAs with suprasellar extension. An APSS ≥25.7 mm is associated with an increased risk of incomplete resection and may outperform conventional grading systems in preoperative planning. Pituitary adenomas (PAs) (Pituitary neuroendocrine tumors) are among the most common intracranial neoplasms, with an estimated prevalence approaching 20% of the population.[ 1 ] Endoscopic endonasal transsphenoidal surgery has become the treatment of choice for most patients and achieves superior rates of gross total resection (GTR), hormonal remission in functioning subtypes, and improved quality of life compared with microscopic approaches.[ 2 - 4 ] Despite these advances, complete resection is not consistently achievable, particularly in tumors that extend beyond the sella. Parasellar invasion, especially into the cavernous sinus, has been extensively studied and remains the best-recognized barrier to GTR.[ 5 , 6 ] However, suprasellar extension introduces distinct anatomical and technical limitations. Tumor growth above the planum sphenoidale narrows the surgical angle, risks compression of the optic apparatus, and may be obscured intraoperatively by descent of the diaphragma sellae during debulking.[ 7 , 13 ] These factors contribute to lower resection rates and higher complication risks in this subgroup, yet current grading systems do not adequately capture this dimension of complexity. The Knosp classification, widely used for assessing cavernous sinus invasion, is a qualitative, coronal-based scale subject to interobserver variability.[ 8 - 11 ] The Zurich Pituitary Score (ZPS) provides a quantitative alternative and has undergone multicenter validation.[ 12 ] Both scales, however, were designed to evaluate parasellar growth and do not address the geometrical constraints of suprasellar disease. As a result, their predictive accuracy diminishes in adenomas whose dominant growth is vertical and sagittal. Recent evidence underscores the importance of sagittal morphometrics in predicting surgical outcomes. Analyses of large and giant adenomas demonstrate that residual disease frequently persists in the suprasellar compartment, particularly along the anterior–posterior axis.[ 14 , 15 ] Park et al. showed that the anteroposterior (AP) dimension is a key determinant of resectability and complications in adenomas with suprasellar extension.[ 16 ] Similarly, novel predictive models, such as the HACKD score, have begun to incorporate AP thresholds as critical parameters.[ 19 ] Yet no suprasellar-specific predictor has been validated for routine clinical use. We hypothesized that the anteroposterior suprasellar diameter (APSS), a simple sagittal measurement obtained on standard magnetic resonance imaging (MRI) above the planum sphenoidale, could serve as a reproducible and clinically relevant predictor of surgical resectability. Unlike coronal-based scales, APSS directly reflects the depth of tumor that lies beyond the working angle of the endonasal corridor. A validated cutoff could assist surgeons in preoperative planning, guide decisions regarding extended approaches, and provide patients with realistic expectations regarding the likelihood of complete resection and need for adjuvant therapy. To test this hypothesis, we conducted a retrospective cohort study of consecutive patients undergoing endoscopic endonasal surgery for PAs with suprasellar extension. Specifically, we reassessed the predictive value of Knosp and ZPS in this anatomical context and systematically evaluated suprasellar morphometrics focusing on APSS as independent predictors of GTR. We conducted a retrospective cohort study at a tertiary neurosurgical referral center. Consecutive adult patients who underwent endoscopic endonasal surgery for PAs with suprasellar extension between July 2021 and June 2024 were included. The study protocol was approved by the institutional bioethics committee and adhered to the principles of the Declaration of Helsinki. Inclusion criteria were: Preoperative MRI demonstrating a sellar mass with suprasellar extension >10 mm above the planum sphenoidale Histopathological confirmation of PA Availability of early postoperative MRI (within 72 h after surgery). Exclusion criteria were: Previous pituitary surgery Pituitary apoplexy at presentation Nonadenomatous histology Incomplete or missing imaging data. All operations were performed using a binostril endoscopic endonasal approach with a 0° lens. Tumor removal combined intracapsular and extracapsular dissection with ring curettes under direct visualization. Cytoreduction proceeded until no residual tumor was visible within the operative field or beyond the endoscope’s working angle. For dural closure, a standardized multilayer technique was employed. Small arachnoid openings were sealed with a synthetic collagen graft (Duragen®, Integra LifeSciences) reinforced with fibrin or BioGlue® sealant. For intraoperative cerebrospinal fluid (CSF) leaks or large dural defects, autologous fat and fascia lata grafts were added, occasionally supported by a nasoseptal flap. After sellar exposure, tumor removal proceeded with stepwise internal decompression followed by extracapsular dissection when a definable pseudocapsule was present. The suprasellar component was approached through the diaphragmatic opening using 0° and, when necessary, 30° endoscopy with angled suction and ring curettes. Diaphragmatic descent (dynamic invagination into the sella after debulking) was assessed endoscopically; a persistent superior bulge, nodularity, or lack of expected collapse suggested residual suprasellar tumor beyond the working angle. Diaphragmatic traction maneuvers were not routinely performed; instead, angled visualization and instruments were preferentially used. Conversion to an extended transtuberculum/transplanum approach with selective intradural exploration was reserved for cases in which the suprasellar component was clearly inaccessible through a transsellar corridor without undue traction or risk to the optic apparatus. MRI-based protocol and morphometric evaluation of suprasellar extension MRI studies were acquired preoperatively and within 72 h postoperatively using 1.5T or 3T scanners. Sequences included T1-weighted sagittal and coronal views with and without gadolinium. The following variables were analyzed: Knosp grade (0–4) based on parasellar extension relative to the intracavernous carotid artery ZPS (grades I–IV) reflecting the proportion of carotid encasement Suprasellar morphometrics (all in millimeters), measured according to a standardized sagittal and coronal protocol: [ 14 ] APSS: horizontal diameter above the planum sphenoidale Suprasellar width (ASS): maximal coronal diameter above the planum Suprasellar height (HSS): maximal vertical diameter above the planum Retrosellar extension (ERS): posterior extension beyond the dorsum sellae Maximum working length (Lmax): distance from the rostral tumor margin to its most distant suprasellar boundary. Measurements were independently performed by a senior neurosurgeon and a neuroradiology fellow using WebDiagRX DICOM viewer v3.2.38. In cases of disagreement, a third senior reviewer adjudicated. The primary outcome was the extent of resection on early postoperative MRI, categorized as: GTR: no visible residual tumor Subtotal resection (STR): any detectable residual. Secondary outcomes included postoperative complications (CSF leak, surgical site hematoma, mortality) and need for reoperation or adjuvant therapy. Intraoperative CSF leaks were managed according to flow. Low-flow leaks did not automatically terminate resection if residual tumor remained safely accessible and visualization was maintained. In the setting of a high-flow leak or large arachnoid defect, further suprasellar manipulation was limited when necessary for safety, and reconstruction was escalated using a graded multilayer technique (collagen matrix, autologous fat/fascia lata as needed, and nasoseptal flap for larger defects), reinforced with sealant. Continuous variables were expressed as mean ± standard deviation (SD) or range; categorical variables as frequencies and percentages. Comparisons of GTR rates across dichotomized Knosp (0–2 vs. 3A–4) and ZPS (I–II vs. III–IV) were performed with Fisher’s exact test. Binary logistic regression assessed independent associations between suprasellar morphometrics, Knosp, and ZPS with the extent of resection. Receiver operating characteristic (ROC) analysis was used to identify the optimal APSS threshold using the Youden’s index, with sensitivity, specificity, and area under the curve reported. A two-sided P < 0.05 was considered statistically significant. Analyses were conducted in IBM Statistical Package for Social Sciences Statistics v29.0.2.0. Of 78 screened patients, 48 met inclusion criteria (30 excluded: prior surgery n = 2; nonadenomatous histology n = 5; missing imaging n = 23). Mean age was 52.2 ± 11.6 years; 52.1% male. Histology: nonfunctioning adenomas 81.3%, growth hormone (GH)-secreting 10.4%, prolactinomas 6.3%, adrenocorticotropic hormone (ACTH)-secreting 2.1%. Postoperative complications occurred in 13 patients (CSF leak n = 8; hematoma n = 1; deaths n = 2 within 72 h due to nosocomial complications). Sphenoid sinus types: sellar 41.7%, presellar 33.3%, postsellar 20.8%, conchal 4.2%.[ 16 ] Parasellar invasion ≥3A (Knosp) or ZPS III–IV occurred in >20%. Suprasellar measurements (mm): APSS 23.1 (11.6–42.9); ASS 22.9 (7.8–42.1); HSS 15.6 (10–39.2); ERS 4.45 (0–24.5); Lmax 44.5 (12–56.5) [ Table 1 ]. Demographic, clinical, and imaging characteristics. GTR occurred in 34/48 (70.8%). Knosp (0–2 vs. 3A–4) did not associate with GTR (P = 0.055); ZPS (I–II vs. III–IV) likewise was nonsignificant (P = 0.258). In logistic regression including morphometrics and parasellar scales, APSS was independently associated with resectability (P = 0.023). ROC analysis identified APSS = 25.7 mm as the optimal threshold to predict failure to achieve GTR (sensitivity 78%, specificity 88%); dichotomized testing at this threshold remained significant (P < 0.01). In a descriptive subanalysis of STR cases, 78.6% had APSS ≥25.7 mm, consistent with the ROC-derived sensitivity. Representation across Knosp and ZPS categories in the STR subset highlights the limited applicability of parasellar-based scales in predominantly suprasellar disease. A total of 78 patients met the inclusion criteria; however, 30 were excluded (2 patients had a history of previous surgery, 5 had a histopathological report other than PA, and pre or postoperative imaging could not be retrieved for 23 cases), resulting in a final sample of 48 participants. A summary of the demographic, clinical, and imaging characteristics of the patients is presented in Table 1 . Of the sample, 25 patients were male (52.1%) and 23 were female (47.9%), with an age range of 18–72 years (mean age 52.23 ± 11.55). The majority of histopathological diagnoses were nonfunctioning PAs (81.3%), followed by GH-secreting adenomas (10.4%), prolactinomas (6.3%), and ACTH-secreting adenomas (2.1%). Postoperative complications were documented in 13 participants (8 cases of CSF fistula, 1 postoperative hematoma, and 2 deaths). Both deaths occurred within the first 72 h postoperatively, and control MRIs were performed within this period. The causes of death were healthcare-associated complications: one patient died of nosocomial pneumonia, and the other from septic shock. More than 20% of the study population had advanced parasellar invasion (Knosp grade 3A or higher, ZPS III-IV) at diagnosis. The most common sphenoid sinus pneumatization type in the studied population was the sellar type (41.7%), followed by presellar (33.3%), postsellar (20.8%), and the conchal type (4.2%) as the least common. The average measurements (in millimeters) for suprasellar extension were: APSS 23.12 (range 11.6–42.9), ASS 22.91 (7.8–42.1), HSS 15.63 (10–39.2), ERS 4.45 (0–24.5), and Lmax 44.5 (12–56.5). GTR was achieved in 34 cases (70.8%), while residual tumor was reported in the remaining 14 cases (29.2%) according to postoperative MRI. Table 2 summarizes the bivariate analysis of the relationship between GTR and parasellar extension categories (Knosp and ZPS). For better analysis, both variables were stratified into two categories: those not representing invasion beyond the cavernous carotid artery (Knosp 0–2 and ZPS I-II) and those that do (Knosp 3A-4 and ZPS III-IV). No significant differences were observed in GTR rates between groups within each variable (P = 0.055, P = 0.258). Relationship of parasellar invasion categories with the degree of macroscopic resection. Subsequently, a logistic regression analysis was performed to determine the relationship of independent variables (APSS, ASS, HSS, ERS, Lmax, Knosp 0–2 vs. 3A-4, ZPS I-II vs. IIIIV) with the dependent variable (GTR degree). Results are detailed in Table 3 . Logistic regression analysis. Degree of macroscopic resection in relation to parasellar invasion categories and measures of suprasellar extension. The AP diameter (APSS) was the only predictive variable with statistical significance (P = 0.023). To determine the cutoff point with the highest discriminative ability, a ROC curve [ Figure 1 ] was generated using all APSS values, and the Youden’s index was calculated. A cutoff point of 25.7 mm was identified, showing the highest sensitivity and specificity (S 78%, E 88%) for predicting failure to achieve GTR. To assess the relationship between APSS at the 25.7 mm cutoff and GTR, a bivariate analysis using Fisher’s exact test was conducted, confirming statistical significance (P ≤ 0.01) [ Table 4 ]. Bivariate analysis of the relationship between the degree of macroscopic resection and the anteroposterior suprasellar diameter with a cutoff point of 25.7 mm. Measurements of suprasellar tension, (a) mid-sagittal section. Blue line: Horizontal line parallel to the sphenoidal planum. Red Line: Vertical line perpendicular to the ephenoidal planum with intersection on the sellar dorsum. Green Line: Suprasellar anteroposterior diameter parallel to the axial line. Orange Line: Suprasellar height parallel to the red line. Yellow line: retrosellar extension of the green line after the red line. Pink line: Maximum working length measured from the most rostral point of the tumor to its furthest margin. (b) Coronal cut. Purple line: Maximum horizontal measured suprasellar width above the sphenoid planum. Additionally, a subanalysis was conducted on cases with subtotal gross resection [ Table 5 ]. Of these, 13 cases (92.85%) were nonfunctioning adenomas. Patients from all categories of the Knosp classification were represented, with 100% representation in categories 3B (n = 1) and 4 (n = 2). In the ZPS classification, all cases were grade II or higher, including 100% (n = 2) of cases classified as grade IV. The averages and SDs of the suprasellar extension measurements are shown in Table 5 . Demographic, clinical, and imaging characteristics of the subgroup of cases with subtotal macroscopic resection. In 11 cases (78.6%), the APSS was equal to or >25.7 mL, aligning with the sensitivity obtained from the ROC curve for APSS measurements. This secondary group included the two patients who died in the postoperative period. Of the remaining patients, 50% underwent reoperation for residual tumor resection at a later time. Three were referred for radiotherapy, and two were followed up without further intervention (one declined reoperation, and the other was managed with octreotide due to a GH-secreting adenoma).","thumbnailUrl":"https://sni-digital-videos.s3.amazonaws.com/articles/c26a5633-a8a9-48a3-a59a-bfbcdab19d08/featured/hero-1781557642357.png","publishDate":"2026-04-17T00:00:00.000Z","doi":"10.25259/SNI_1233_2025","categories":["Neuro-oncology","Original Article"],"fullTextUrl":"https://surgicalneurologyint.com/api/articles/c26a5633-a8a9-48a3-a59a-bfbcdab19d08/pdf"}