General discussion 507 13 oncological diagnostics. MD may distinguish neoplasms by their tumour biology and behaviour. As such, molecular risk-stratification may provide prognostic and potentially therapeutic information that morphology alone may not and that is beyond the appreciation of test performance parameters such as sensitivity and specificity. Moreover, despite any limitations regarding potential falsenegative results as described above, MD may guarantee a more objective categorization, whereas intra- and interobserver variation will always plague the morphological assessment of cytology as well as histopathology [26, 721, 722, 749, 750]. For example, in oncocytic nodules, long-term metastases of morphologically benign lesions are notorious. In retrospect, the initial index nodule had a malignant molecular profile and may initially better have been managed as malignancy [695]. Although the morphological histopathological assessment cannot be renounced, we foresee that MD-based diagnosis of cytology and histopathology may soon take over from morphological assessment as the primary cytological classifier [751], as the determinant for treatment decisions (e.g., no surgery, thyroid lobectomy or total thyroidectomy), and as (part of) the reference standard for histopathology. In a way, we already see the transition to an MD-based diagnosis in our daily practice as well as in the EfFECTS trial, for example regarding the two [18F]FDG (false-)negative nodules described above that were only classified as malignant following the MD results [501, 725]. We also hypothesize that the prognosis of differentiated thyroid carcinoma may shift when neoplasms are categorized and managed according to their molecular profile. With the overall excellent prognosis of differentiated thyroid carcinoma, however, any changes in prognosis will be subtle and will take large and long-term prognostic studies to demonstrate. MD-based classification of thyroid nodules may change the interpretation of the results of the EfFECTS trial, too. For example, as presented in the supplementary data of Chapter 12, there were several cases (cases 114, 123, and 127-130, supplementary table 5) in which both MD and [18F]FDGPET/CT were considered false-positive based on blinded morphological assessment as the reference standard (i.e., without knowledge of MD and [18F]FDG-PET/CT results). However, these cases have an undeniable malignant molecular profile. According to an MD-based diagnosis, these nodules would be considered malignant and both tests would be considered true-positive. The number of false-negative [18F]FDG-PET/CT results would change as well. When an MD-based diagnosis would be applied to distinguish any nodules that require surgical resection (i.e., borderline/premalignant and malignant) from benign nodules based on an uncertain malignant or malignant molecular diagnosis (supplementary table 5 of Chapter 12 of this thesis), seven cases that are now considered true-negative on [18F]FDG-PET/CT (case 3-9, supplementary table 5) would have to be reconsidered as false-negative. Still, their mere intermediate risk molecular alterations and [18F]FDG negativity likely indicate biologically indolent behaviour and it may be contestable how “false” negative these nodules really are. The biological truth may be somewhere in between.
RkJQdWJsaXNoZXIy MTk4NDMw