Is it Time for a New Approach to the Differential Diagnosis of ACTH-Dependent Cushing Syndrome?

Thirty years ago, Oldfield’s seminal publication on inferior petrosal sinus sampling (IPSS) (1) changed our approach to the differential diagnosis of adrenocorticotropic hormone (ACTH)-dependent Cushing syndrome (CS). This year, Frete et al may have accomplished the same feat by suggesting that a new diagnostic strategy reduces the need for IPSS (2). To understand the impact of this proposal, we must consider the underlying assumptions, evolution, and use of diagnostic tests.

All tests carry implicit assumptions. When using IPSS, for example, we assume that blood is drawn and processed correctly, the ACTH assay is precise and accurate, and the radiologist is experienced. However, unforeseen issues reduced 100% diagnostic accuracy to about 95%: it became clear that patients with cyclic hypercortisolism and ectopic ACTH syndrome (EAS) responded to corticotropin-releasing hormone (CRH) (and IPSS) if normal corticotropes were not suppressed by consistent hypercortisolism, leading to a false diagnosis of Cushing disease (CD). Conversely, arborized or atrophic petrosal sinuses were associated with a lack of a central step-up in ACTH, resulting in a false EAS diagnosis (3).

Tests are modified when clinician investigators discover the reason(s) for test failure. This may occur when the clinician initially acts on the test and later discovers the diagnostic error. Alternatively, an experienced clinician may intuit that the patient’s presentation does not fit the test’s results. When no physiologic or technical reason explains test failure, the criterion for interpretation may need to change. Ultimately, tests may be abandoned because of diagnostic inaccuracy, adverse events, high cost, or lack of availability.

How do these considerations relate to IPSS? IPSS relies on a straightforward question: Is the pituitary gland the source of ACTH? This elegant strategy, with its high diagnostic accuracy, led to statements that it could (or should) serve as the only test for the differential diagnosis of ACTH-dependent CS when an unequivocal lesion is not seen on pituitary magnetic resonance imaging (MRI) (4).

Thirty years of experience now show that optimal results require extensive experience about when and how to perform and interpret the test, implying that IPSS should be performed only in high-volume centers. Furthermore, despite the discovery that desmopressin could be substituted for CRH, some countries still cannot access either agent. Additionally, new algorithms for pituitary MRI, increased computed tomography (CT) resolution, and functional imaging with somatostatin analogs and ¹⁸F-DOPA resulted in dramatic improvements in tumor detection, suggesting that imaging might be a more direct diagnostic approach. Thus, it is now timely to look for approaches that are equally useful and more available than IPSS.

The diagnostic strategy advanced in the Frete paper exploits state-of-the-art imaging techniques and embodies best practices of interdisciplinary collaboration. In this study of 167 CD patients and 27 EAS patients, patients underwent pituitary MRI and desmopressin (DDAVP) and CRH tests; IPSS was optional (2). If MRI showed a lesion and both tests indicated CD, the patient could proceed to transsphenoidal surgery. All other results were followed by triple-phase CT, and possibly IPSS if no lesion was seen. The combination of noninvasive tests (DDAVP, CRH, MRI. and possibly CT) gave a correct diagnosis in 53% of patients, potentially halving the need for IPSS to achieve high diagnostic certainty.

What does this paper teach regarding best practices and limitations? Most importantly, how should it change our diagnostic approach?

For Cushing syndrome, diagnostic acumen is more than interpretation of a single dynamic endocrine test—it involves analysis of imaging, synthesis of all results and conceptualization of the pre- and post-test(s) probability of the cause. As in this study, a multidisciplinary team of endocrinologists, surgeons, and radiologists is essential to achieve consensus management decisions through vigorous discussions about all aspects of the history, physical findings, and test results. This strategy facilitates inductive thinking and examination of bias in diagnostic heuristics, and minimizes continuity and anchoring bias (failure to discard an initial presumed diagnosis) (5).

Additional best practices included inclusion of sufficient EAS patients for comparison to CD, and surgical confirmation of the diagnosis in nearly all patients. The two-center design reduces concerns about generalizability to similar academic centers. Importantly, the 15mSV effective dose for neck-to-pelvis CT reduces concern about radiation exposure, as compared with 16.7 ± 8.9 (SD) mSV for abdomen/chest CT in the US (6). Also, the 2-mm slice thickness should increase tumor detection.

Operational limitations may reduce the availability of this approach. An experienced multidisciplinary team and a pituitary surgeon with a near 100% success rate were necessary for optimal results. There is an assumption that imaging, performed as described, will detect all cases of EAS. However, many EAS lesions are smaller than 10 mm; CT detected only 4/6 of these. Although CT identified the tumor in 2 EAS patients falsely categorized by endocrine tests, absent such detection, they would have had unnecessary pituitary surgery.

Additional development of the strategy should be considered:

• The criteria for CRH and DDAVP test identification of CD are lower than those recommended by some—favoring increased sensitivity (7). The 74% sensitivity needs additional validation. If it deteriorates, or continues to falsely diagnose EAS, the algorithm will require adjustment.

• This dual test approach may work when substituting high-dose dexamethasone for DDAVP, improving global availability. When Reimondo et al tested this, IPSS could be avoided in 53%, similar to the current study (47%) (8).

• Diagnostic criteria for the tests were not optimized for fewer time-points to reduce cost.

• Because there were no EAS tumors that are typically harder to identify, such as gut and prostatic NETs, generalizability to all EAS patients is unknown.

• EAS tumor detection might improve with SSTR analogs other than octreotide.

How should we incorporate these results into practice? CS patients should be evaluated at centers with the requisite interdisciplinary expertise and IPSS and imaging capability to consider validation of this approach; ideally, multi-site collaborations would foster analysis of a larger population.

Frete et al have pushed us past a continuation bias related to current diagnostic recommendations; I look forward to vigorous conversations and exploration of this new approach.

Glossary

Abbreviations

ACTH

adrenocorticotropic hormone

CRH

corticotropin-releasing hormone

CT

computed tomography

DDAVP

desmopressin

EAS

ectopic ACTH syndrome

IPSS

inferior petrosal sinus sampling

MRI

magnetic resonance imaging

Additional Information

Disclosure Summary: Dr Nieman is an author of chapters in UpToDate on Cushing’s syndrome and receives royalties for that work.

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.