General discussion 513 13 Challenges in diagnostic trials Progressive diagnostic insights: overtaken by time? Counting from the first approval of the trial protocol on 10 November 2014, local procedures to initiate the 15 study sites took anywhere between 7 and 24 months. Counting from the date of the inclusion of the first patient on 16 July 2015, recruitment of the 132 participants of the EfFECTS trial took exactly three years and three months. Follow-up of all patients was completed on 14 November 2019, the date that the final follow-up ultrasound was performed for the last patient undergoing active surveillance for an [18F]FDG negative nodule. With the intervention of a pandemic, data collection and analysis took another two years to complete. Finally, the paper with the main results of the EfFECTS trial was accepted for publication on 16 November 2021, seven years after the trial protocol was approved [501]. Even though a timeframe like this is not unusual in clinical trials, it shows one of the difficulties in clinical research: time - and thereby progress - goes on. During the conduction of the EfFECTS trial, the differentiation of thyroid nodules prior to FNAC using ultrasound classification systems such as the ATA ultrasound classification and various versions of the TIRADS classification was gaining momentum [17, 394, 395, 488]. TIRADS is now integrated in our daily practice to select only ultrasonographically suspicious nodules for FNAC. Consequently, fewer FNACs are performed and the a priori malignancy risk of cytology is likely higher, as the consecutive versions of the Bethesda classification also suggest [16, 18, 27]. This higher prevalence of malignancy puts higher demands on the diagnostic accuracy of [18F]FDG-PET/CT to meet the ATA requirements NPV of an ‘accurate rule-out test’ [17]. As TIRADS did not gain significant traction in the Netherlands until the majority of patients was included in the EfFECTS trial, the description of a TIRADS classification was not part of the trial protocol and not available in the baseline ultrasound reports of most patients. These baseline ultrasounds were performed as a part of clinical practice, prior to FNAC and study inclusion. Moreover, as ultrasound is a dynamic technique, we considered it inappropriate to re-evaluate the archived ultrasound images and retrospectively reassess them according to TIRADS for the papers with the main trial results [501]. Although methodologically we still support our choices, we have encountered significant criticism from multiple peer-reviewers during the review process of several of our papers. After all, using TIRADS, the altered preselection of patients may also tamper with the diagnostic accuracy, costeffectiveness and overall clinical utility of [18F]FDG-PET/CT that we established in the EfFECTS trial, for better or for worse. This is nowadays considered an important limitation of our work and may not give the results of the EfFECTS trial as much weight as they might deserve [501, 725, 726].
RkJQdWJsaXNoZXIy MTk4NDMw