63 To pet or not to pet?. (20th August 2018)

• Record Type: Journal Article
• Authors: Badea, Andreea; Vanderby, Sonia; Babyn, Paul
• Is Part Of: BMJ evidence-based medicine. Volume 23:Supplement 2(2018)
• Page Start: A29
• Page End: A30
• Publication Date: 2018-08-20
• Subjects: Evidence-based medicine -- Periodicals; 616.005
• DOI: 10.1136/bmjebm-2018-111070.63
• Languages: English
• ISSNs: 2515-446X

Abstract

Abstract : Objectives: Concerns about unnecessary medical imaging have led to the development of guidelines for a variety of indications and modalities by radiology and specialty groups. This study evaluates the utility of four national imaging guidelines, which assess the value of positron emitting tomography/computed tomography (PET/CT). As PET/CT is an emerging technology, the evidence base for the development of guidelines is not as established as other modalities. Of the indications commonly explored by PET/CT, the evaluation of lung cancer and lung nodules has the greatest evidence base and is therefore the most evaluable sector for guideline comparison. The primary goal of this study is to determine which of the four guidelines has the greatest coverage in a real-world clinical setting. The secondary goal of this study was to assess the convenience of the utilization of the selected guidelines for medical imaging specialists in a high-volume practice setting. Method: PET/CT exam requests for suspected or confirmed lung cancer and lung nodules within a specified common seven day period were retrospectively accessed from four centers in Canada. Requests were de-identified and combined to create a single data set. Requests were then evaluated using four sets of recognized radiology guidelines: (a) the Canadian Association of Radiologists Referral Guidelines (CAR-RG), (b) the American Association of Radiologists Appropriateness Criteria (ACR-AC), (c) National Comprehensive Cancer Network (NCCN) Practice Guidelines, and (d) the Institut National d'Excellence en Santé et en Services Sociaux (INESSS) Interactive Tool. Results were compared in terms of (i) condition/symptom coverage and (ii) recommendations obtained with particular attention to conflicting or differing results. Rates of high-value, low-value and unscorable requests were tabulated and rates were compared among guidelines. As a secondary assessment of the coverage, the usability of each set was subjectively measured. Results: In total, 74 requests qualified for the inclusion criteria. Of these, 51% were requested for confirmed or highly suspicious cases of lung cancer and 49% were for the evaluation of lung nodules. 36%–57% of all exams evaluated were considered high-value and 3%–7% were considered low-value. Requests that could not be assessed totaled 31%–50% for lack of information and 1%–9% for the lack of an applicable guideline. Guideline coverage ranged from 64%–72%. Of the guidelines used, the INESSS Interactive Tool was the most user-friendly in terms of organization and information provided with the value assessment. While the ACR-AC are very well developed for the radiologic management of cancers, the guidelines for their imaging was more sparse and difficult to navigate. The NCCN guidelines provide an easy to use summary document. The CAR-RG while well organized by body system, provided vague guidelines. Conclusions: While the choice of guidelines used by institutions is generally determined based on the country they are in, our research has demonstrated that in some instances, it is prudent to use the guidelines that are most developed for that indication. For lung cancer cases and lung nodule evaluation, the INESSS guidelines performed the best with only 28% of total requisitions unable to be scored, whereas the CAR-RG guidelines cover the least indications with 36% of requests unable to be assessed. These results also indicate areas where national guidelines would benefit from further development. It is important to note that the INESSS guidelines also resulted in the most high-value assessments at 42% of requests. This might suggest that the issue does not always lie in low-value ordering, but perhaps the lack of guideline coverage and clarity in some guidelines resulting in higher proportions of low-value scores.