Machine learning-based prediction of outcomes of the endoscopic endonasal approach in Cushing disease: is the future coming?

OBJECTIVE

Machine learning (ML) is an innovative method to analyze large and complex data sets. The aim of this study was to evaluate the use of ML to identify predictors of early postsurgical and long-term outcomes in patients treated for Cushing disease (CD).

METHODS

All consecutive patients in our center who underwent surgery for CD through the endoscopic endonasal approach were retrospectively reviewed. Study endpoints were gross-tumor removal (GTR), postsurgical remission, and long-term control of disease. Several demographic, radiological, and histological factors were assessed as potential predictors. For ML-based modeling, data were randomly divided into 2 sets with an 80% to 20% ratio for bootstrapped training and testing, respectively. Several algorithms were tested and tuned for the area under the curve (AUC).

RESULTS

The study included 151 patients. GTR was achieved in 137 patients (91%), and postsurgical hypersecretion remission was achieved in 133 patients (88%). At last follow-up, 116 patients (77%) were still in remission after surgery and in 21 patients (14%), CD was controlled with complementary treatment (overall, of 131 cases, 87% were under control at follow-up). At internal validation, the endpoints were predicted with AUCs of 0.81-1.00, accuracy of 81%-100%, and Brier scores of 0.035-0.151. Tumor size and invasiveness and histological confirmation of adrenocorticotropic hormone (ACTH)-secreting cells were the main predictors for the 3 endpoints of interest.

CONCLUSIONS

ML algorithms were used to train and internally validate robust models for all the endpoints, giving accurate outcome predictions in CD cases. This analytical method seems promising for potentially improving future patient care and counseling; however, careful clinical interpretation of the results remains necessary before any clinical adoption of ML. Moreover, further studies and increased sample sizes are definitely required before the widespread adoption of ML to the study of CD.

Optimizing pre-, intra-, and postoperative management of patients with sellar pathology undergoing transsphenoidal surgery

OBJECTIVE

Perioperative management of patients with sellar lesions is complex, requiring input from a multidisciplinary team of specialists for ongoing management of both endocrinological and neurosurgical issues. Here, the authors reviewed the experience of a single multidisciplinary center over 10 years to identify key postoperative practices that ensure positive outcomes for patients with sellar lesions who undergo transsphenoidal surgery.

METHODS

The authors performed a retrospective review of all transsphenoidal operations carried out by the senior author at a single center from April 2008 through November 2018. They included only adult patients and recorded perioperative management. They also reviewed the evolution of clinical practices for perioperative care at their institution to identify strategies for ensuring positive patient outcomes, and they reviewed the literature on select related topics.

RESULTS

In total, 1023 operations in 928 patients were reviewed. Of these, 712 operations were for pituitary adenomas (69.6%), and 122 were for Rathke cleft cysts (11.9%). The remainder included operations for craniopharyngiomas (3.6%), arachnoid cysts (1.7%), pituitary tumor apoplexy (1.0%), and other sellar pathologies (12.2%). Among the reviewed operations, the median hospital stay was 3 days (IQR 2-3). Patient management details during the pre-, intra-, and postoperative periods were identified, including both shared characteristics of all patients undergoing transsphenoidal surgery and unique characteristics that are specific to certain lesion types or patient populations.

CONCLUSIONS

Patients with sellar lesions who undergo transsphenoidal surgery require complex, multidisciplinary perioperative care to monitor for common adverse events and to improve outcomes, but there is a dearth of high-quality evidence guiding most perioperative practices. Here, the authors reviewed practices at their institution across more than 1000 transsphenoidal operations that may help ensure successful patient outcomes.

Negative surgical exploration in patients with Cushing's disease: benefit of two-thirds gland resection on remission rate and a review of the literature

OBJECTIVEThe authors report their single-institution experience with the pathological findings, rates of remission, and complications in patients with presumed Cushing's disease (CD) who underwent a two-thirds pituitary gland resection when no adenoma was identified at the time of transsphenoidal surgery (TSS). The authors also review the literature on patients with CD, negative surgical exploration, and histological findings.METHODSThis study is a retrospective analysis of cases found in neurosurgery and pathology department databases between 1989 and 2011. In all cases, patients had been operated on by the same neurosurgeon (K.O.L.). Twenty-two (13.6%) of 161 patients who underwent TSS for CD had no adenoma identified intraoperatively after systematic exploration of the entire gland; these patients all underwent a two-thirds pituitary gland resection. A chart review was performed to assess treatment data points as well as clinical and biochemical remission status.RESULTSOf the 22 patients who underwent two-thirds gland resection, 6 (27.3%) ultimately had lesions found on final pathology. All 6 patients were found to have a distinct adrenocorticotropic hormone (ACTH) cell adenoma. Sixteen (72.7%) of the patients had no tumor identified, with 3 of these patients suspected of having ACTH cell hyperplasia. The follow-up duration for the entire group was between 14 and 315 months (mean 98.9 months, median 77 months). Remission rates were 100% (6/6 patients) for the ACTH cell adenoma group and 75% (12/16) for the group without adenoma. Overall, 18 (81.8%) of the 22 patients had no evidence of hypercortisolism at last follow-up, and 4 patients (18%) had persistent hypercortisolism, defined as a postoperative cortisol level > 5 μg/dl. Of these 4 patients, 1 was suspected of harboring a cavernous sinus adenoma, 2 were found to have lung tumors secreting ACTH, and 1 remained with an undiagnosed etiology. Rates of postoperative complications were low.CONCLUSIONSThe diagnosis and treatment of CD can be challenging for neurosurgeons, endocrinologists, and pathologists alike. Failure to find a discrete adenoma at the time of surgery occurs in at least 10%-15% of cases, even in experienced centers. The current literature provides little guidance regarding rational intraoperative approaches in such cases. The authors' experience with 161 patients with CD, when no intraoperative tumor was localized, demonstrates the utility of a two-thirds pituitary gland resection with a novel and effective surgical strategy, as suggested by a high initial remission rate and a low operative morbidity.

Potential utility of FLAIR in MRI-negative Cushing's disease

OBJECTIVE Accurate presurgical localization of microadenomas in Cushing's disease (CD) leads to improved remission rates and decreased adverse events. Volumetric gradient recalled echo (3D-GRE) MRI detects pituitary microadenomas in CD in up to 50%-80% cases as a focus of hypointensity due to delayed contrast wash-in. The authors have previously reported that postcontrast FLAIR imaging may be useful in detecting otherwise MRI-negative pituitary microadenomas as foci of hyperintensity. This reflects theoretically complementary imaging of microadenomas due to delayed contrast washout. The authors report on the diagnostic accuracy and clinical utility of FLAIR imaging in the detection of microadenomas in patients with CD. METHODS The authors prospectively analyzed imaging findings in 23 patients (24 tumors) with biochemically proven CD who underwent transsphenoidal surgery for CD. Preoperatively, the patients underwent pituitary MRI with postcontrast FLAIR and postcontrast 3D-GRE sequences. RESULTS Postcontrast FLAIR hyperintensity was detected in macroadenomas, and in 3D-GRE-positive or -negative microadenomas. Overall, 3D-GRE was superior in detecting surgically and histopathologically confirmed, location-concordant microadenomas. Of 24 pituitary adenomas, 18 (75%; sensitivity 82%, positive predictive value 95%) were found on 3D-GRE, and 13 (50% [1 was false positive]; sensitivity 55%, positive predictive value 92%) were correctly identified on FLAIR. The stand-alone specificity of 3D-GRE and FLAIR was similar (50%). These results confirm the superiority of 3D-GRE as a stand-alone imaging modality. The authors then tested the utility of FLAIR as a complementary tool to 3D-GRE imaging. All 5 patients with negative 3D-GRE MRI displayed a distinct focus of FLAIR enhancement. Four of those 5 cases (80%) had location-concordant positive histopathological results and achieved postsurgical biochemical remission. The remaining patient was not cured, because resection did not include the region of FLAIR hyperintensity. CONCLUSIONS This study suggests that delayed microadenoma contrast washout may be detected as FLAIR hyperintensity in otherwise MRI-negative CD cases. The authors propose adding postcontrast FLAIR sequences to complement 3D-GRE for surgical planning in patients with CD. Clinical trial registration no.: NIH protocol 03-N-0164, NCT00060541 (clinicaltrials.gov).

Comparison of MRI techniques for detecting microadenomas in Cushing's disease

OBJECTIVE Many centers use conventional and dynamic contrast-enhanced MRI (DMRI) sequences in patients with Cushing's disease. The authors assessed the utility of the 3D volumetric interpolated breath-hold examination, a spoiled-gradient echo 3D T1 sequence (SGE) characterized by superior soft tissue contrast and improved resolution, compared with DMRI and conventional MRI (CMRI) for detecting microadenomas in patients with Cushing's disease. METHODS This study was a blinded assessment of pituitary MRI in patients with proven Cushing's disease. Fifty-seven patients who had undergone surgery for Cushing's disease (10 male, 47 female; age range 13-69 years), whose surgical findings were considered to represent a microadenoma, and who had been examined with all 3 imaging techniques were included. Thus, selection emphasized patients with prior negative or equivocal MRI on referral. The MRI annotations were anonymized and 4 separate imaging sets were independently read by 3 blinded, experienced clinicians: a neuroradiologist and 2 pituitary surgeons. RESULTS Forty-eight surgical specimens contained an adenoma (46 ACTH-staining adenomas, 1 prolactinoma, and 1 nonfunctioning microadenoma). DMRI detected 5 adenomas that were not evident on CMRI, SGE detected 8 adenomas not evident on CMRI, including 3 that were not evident on DMRI. One adenoma was detected on DMRI that was not detected on SGE. McNemar's test for efficacy between the different MRI sets for tumor detection showed that the addition of SGE to CMRI increased the number of tumors detected from 18 to 26 (p = 0.02) based on agreement of at least 2 of 3 readers. CONCLUSIONS SGE shows higher sensitivity than DMRI for detecting and localizing pituitary microadenomas, although rarely an adenoma is detected exclusively by DMRI. SGE should be part of the standard MRI protocol for patients with Cushing's disease.

Cushing's disease: pathobiology, diagnosis, and management

Cushing's disease (CD) is the result of excess secretion of adrenocorticotropic hormone (ACTH) by a benign monoclonal pituitary adenoma. The excessive secretion of ACTH stimulates secretion of cortisol by the adrenal glands, resulting in supraphysiological levels of circulating cortisol. The pathophysiological levels of cortisol are associated with hypertension, diabetes, obesity, and early death. Successful resection of the CD-associated ACTH-secreting pituitary adenoma is the treatment of choice and results in immediate biochemical remission with preservation of pituitary function. Accurate and early identification of CD is critical for effective surgical management and optimal prognosis. The authors review the current pathophysiological principles, diagnostic methods, and management of CD.

Pitfalls in the diagnosis and management of Cushing's syndrome

Despite many recent advances, the management of patients with Cushing's disease continues to be challenging. Cushing's syndrome is a complex metabolic disorder that is a result of excess glucocorticoids. Excluding the exogenous causes, adrenocorticotropic hormone-secreting pituitary adenomas account for nearly 70% of all cases of Cushing's syndrome. The suspicion, diagnosis, and differential diagnosis require a logical systematic approach with attention paid to key details at each investigational step. A diagnosis of endogenous Cushing's syndrome is usually suspected in patients with clinical symptoms and confirmed by using multiple biochemical tests. Each of the biochemical tests used to establish the diagnosis has limitations that need to be considered for proper interpretation. Although some tests determine the total daily urinary excretion of cortisol, many others rely on measurements of serum cortisol at baseline and after stimulation (e.g., after corticotropin-releasing hormone) or suppression (e.g., dexamethasone) with agents that influence the hypothalamic-pituitary-adrenal axis. Other tests (e.g., measurements of late-night salivary cortisol concentration) rely on alterations in the diurnal rhythm of cortisol secretion. Because more than 90% of the cortisol in the circulation is protein bound, any alteration in the binding proteins (transcortin and albumin) will automatically influence the measured level and confound the interpretation of stimulation and suppression data, which are the basis for establishing the diagnosis of Cushing's syndrome. Although measuring late-night salivary cortisol seems to be an excellent initial test for hypercortisolism, it may be confounded by poor sampling methods and contamination. Measurements of 24-hour urinary free-cortisol excretion could be misleading in the presence of some pathological and physiological conditions. Dexamethasone suppression tests can be affected by illnesses that alter the absorption of the drug (e.g., malabsorption, celiac disease) and by the concurrent use of medications that interfere with its metabolism (e.g., inducers and inhibitors of the P450 enzyme system). In this review, the authors aim to review the pitfalls commonly encountered in the workup of patients suspected to have hypercortisolism. The optimal diagnosis and therapy for patients with Cushing's disease require the thorough and close coordination and involvement of all members of the management team.

Cyclic Cushing's disease with misleading inferior petrosal sinus sampling results during a trough phase

Diagnosing Cushing's syndrome is challenging and is further hampered when investigations are performed in a patient with cyclic Cushing's syndrome. A subset of patients with Cushing's syndrome exhibit periods of abnormal cortisol secretion with interspersed normal secretion. Patients can have periods of clinical improvement during these quiescent phases or remain symptomatic. Initial diagnostic testing can be challenging because of the unpredictable durations of the peak and trough phases, and it is especially challenging when the diagnosis of cyclic Cushing's syndrome has not yet been determined. Here, the authors present the case of a patient with Cushing's disease with a pathology-proven adrenocorticotropic hormone (ACTH)-secreting pituitary adenoma and whose initial inferior petrosal sinus sampling (IPSS) results were deemed indeterminate; further studies elucidated the diagnosis of cyclic Cushing's syndrome. Repeat IPSS was diagnostic of a central source for ACTH secretion, and the patient was treated successfully with transsphenoidal resection. Literature concerning the diagnosis and management of cyclic Cushing's syndrome is also reviewed.

High-resolution(18)F-fluorodeoxyglucose positron emission tomography and magnetic resonance imaging for pituitary adenoma detection in Cushing disease

OBJECT High-resolution PET (hrPET) performed using a high-resolution research tomograph is reported as having a resolution of 2 mm and could be used to detect corticotroph adenomas through uptake of(18)F-fluorodeoxyglucose ((18)F-FDG). To determine the sensitivity of this imaging modality, the authors compared(18)F-FDG hrPET and MRI detection of pituitary adenomas in Cushing disease (CD). METHODS Consecutive patients with CD who underwent preoperative(18)F-FDG hrPET and MRI (spin echo [SE] and spoiled gradient recalled [SPGR] sequences) were prospectively analyzed. Standardized uptake values (SUVs) were calculated from hrPET and were compared with MRI findings. Imaging findings were correlated to operative and histological findings. RESULTS Ten patients (7 females and 3 males) were included (mean age 30.8 ± 19.3 years; range 11-59 years). MRI revealed a pituitary adenoma in 4 patients (40% of patients) on SE and 7 patients (70%) on SPGR sequences.(18)F-FDG hrPET demonstrated increased(18)F-FDG uptake consistent with an adenoma in 4 patients (40%; adenoma size range 3-14 mm). Maximum SUV was significantly higher for(18)F-FDG hrPET-positive tumors (difference = 5.1, 95% CI 2.1-8.1; p = 0.004) than for(18)F-FDG hrPET-negative tumors.(18)F-FDG hrPET positivity was not associated with tumor volume (p = 0.2) or dural invasion (p = 0.5). Midnight and morning ACTH levels were associated with(18)F-FDG hrPET positivity (p = 0.01 and 0.04, respectively) and correlated with the maximum SUV (R = 0.9; p = 0.001) and average SUV (R = 0.8; p = 0.01). All(18)F-FDG hrPET-positive adenomas had a less than a 180% ACTH increase and(18)F-FDG hrPET-negative adenomas had a greater than 180% ACTH increase after CRH stimulation (p = 0.03). Three adenomas were detected on SPGR MRI sequences that were not detected by(18)F-FDG hrPET imaging. Two adenomas not detected on SE (but no adenomas not detected on SPGR) were detected on(18)F-FDG hrPET. CONCLUSIONS While(18)F-FDG hrPET imaging can detect small functioning corticotroph adenomas and is more sensitive than SE MRI, SPGR MRI is more sensitive than(18)F-FDG hrPET and SE MRI in the detection of CD-associated pituitary adenomas. Response to CRH stimulation can predict(18)F-FDG hrPET-positive adenomas in CD.