Macroscopic Fat in Adrenocortical Carcinoma: A Systematic Review
Damithri S. Ranathunga1 Lindsay A. Cherpak1 Nicola Schieda1 Trevor A. Flood2 Matthew D. F. McInnes1,3,4
Keywords: adrenal gland, adrenocortical carcinoma, fat, systematic review
doi.org/10.2214/AJR.19.21851
Received June 11, 2019; accepted after revision July 28, 2019.
Supported by the University of Ottawa Department of Radiology Research Stipend Program.
1Department of Radiology, University of Ottawa, Ottawa, ON, Canada.
2Department of Pathology and Laboratory Medicine, University of Ottawa, Ottawa, ON, Canada.
3Department of Diagnostic Radiology, Ottawa Hospital Research Institute, The Ottawa Hospital, Civic Campus, 1053 Carling Ave, Ottawa, ON K1E 4Y9, Canada. Address correspondence to M. D. F. McInnes (mmcinnes@toh.ca).
4Department of Medical Imaging, The Ottawa Hospital, Ottawa, ON, Canada.
AJR 2020; 214:1-5
0361-803X/20/2142-1
@ American Roentgen Ray Society
OBJECTIVE. The objective of this systematic review was to determine the number and quality of reports of adrenocortical carcinoma (ACC) containing macroscopic fat; this infor- mation may inform guidelines for diagnosis and management of ACC.
MATERIALS AND METHODS. A comprehensive search of databases of published studies was performed. Two reviewers independently selected original research, case series, or case reports of ACC with macroscopic fat on imaging and extracted data. Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool.
RESULTS. Three case reports and one retrospective study comprising a total of seven cases of ACC (lesion size: range, 6.5-22 cm) with macroscopic fat were included. ACC was symptomatic in all patients; neither locally invasive features nor metastases were reported. Four cases had less than 5% macroscopic fat on imaging, and the percentage fat on imaging was not reported for the remaining three cases. With regard to the risk of bias, one case had high risk for the index test domain because of potentially unreliable determination of macro- scopic fat (i.e., no pathologic confirmation). All seven cases (from four studies) had unclear risk for the reference standard domain because there was insufficient information about the reference standard to determine whether ACC was correctly diagnosed. All studies were at low risk of bias in the flow and timing domain.
CONCLUSION. There are few reports of macroscopic fat in ACC detected on imaging studies; among the reports of macroscopic fat in ACC, tumors were large (> 6 cm) and had a small proportion of gross fat (< 5%). The reliability of reported cases is questionable primar- ily because of insufficient details about pathologic diagnosis. Based on this information, a change in guideline recommendations may not be warranted. However, consideration of fol- low-up or biopsy of patients with large symptomatic tumors (> 6 cm) containing a small pro- portion of fat (<5%) may be appropriate.
A drenocortical carcinoma (ACC) is a rare malignancy with an in- cidence of 0.7-2.0 cases per 1 million persons per year. ACC can occur at any age with the first peak inci- dence in the 1st decade and the second peak incidence in patients between 40 and 50 years old with a slight female predominance [1-3]. ACCs are typically sporadic but can be associated with multiple endocrine neoplasia type 1, Lynch syndrome, and Li-Fraumeni cancer syndrome [4].
Two major staging systems used for adre- nal cancer are the American Joint Commit- tee on Cancer TNM staging system and the European Network for the Study of Adrenal Tumors staging system [5]. Stage I tumor is 5 cm or smaller and confined to the adrenal gland. Stage II tumor is larger than 5 cm and
confined to the adrenal gland. Stage III tu- mor has spread to nearby lymph nodes or has invaded beyond the adrenal gland. Stage IV cancer has metastasized [5]. Five-year stage- dependent survival is 66-82% for stage I, 58-64% for stage II, 24-50% for stage III, and 0-17% for stage IV [3, 6].
Because of capacious nature of the retro- peritoneum, most ACCs are not diagnosed until they are large or invasive; thus, stage III or IV disease is diagnosed at presenta- tion in approximately 70% of patients with ACC [6]. ACCs identified earlier are as- sociated with better outcomes. Stage I, II, and occasionally III ACCs are amendable to radical resection, which aims for com- plete surgical resection with curative intent, whereas advanced ACC tumors (stage IV and most stage III tumors) are typically in-
operable and are associated with poor out- comes [3].
The Weiss score is the best validated and most commonly used method to microscopi- cally differentiate between benign and ma- lignant ACC tumors. The nine histologic criteria are high nuclear grade, mitotic rate greater than 5 per 50 HPF, atypical mitotic figures, clear cells comprising 25% or less of tumor, diffuse architecture (> 33% of tumor), necrosis, venous invasion, sinusoidal inva- sion, and capsular invasion. The presence of one item is equal to 1 point, and the sum of the positive items defines the final score. A Weiss score of 3 or greater is diagnostic of an ACC, and a Weiss score of 2 or less is diag- nostic of an adrenal adenoma [7, 8].
ACCs are generally discovered on imag- ing as large heterogeneous, locally aggres- sive adrenal masses [9]. CT and MRI are the main diagnostic imaging modalities to eval- uate ACC; however, there are no specific im- aging findings that can establish the diagno- sis of ACC [9]. Rather, the role of imaging is to stage suspected ACC or to exclude ACC by diagnosis of a benign adrenal mass (e.g., adenoma or myelolipoma) [9]. Both adrenal adenomas and ACCs may show microscopic fat on imaging, because of the presence of fat molecules within the cytoplasm of adre- nal cortical cells [10]. Therefore, one can- not rely on the presence of microscopic fat on imaging alone to differentiate adenoma from ACC; however, this distinction is rarely problematic because adenomas are generally incidentally discovered, small (< 4 cm), ho- mogeneous, and nonaggressive and generally remain stable or grow slowly over time [11].
The presence of macroscopic fat, an imag- ing finding related to the presence of intratu- moral adipocytes (fat cells), is a more contro- versial imaging finding. Macroscopic fat in an adrenal mass is generally considered diag- nostic of adrenal myelolipomas [9]; however, there are sporadic case reports of adrenocor- tical neoplasms, including ACC, also con- taining macroscopic fat [12-15]. For this rea- son, authors have suggested that a definitive diagnosis of myelolipoma not be suggested on imaging unless an adrenal mass contains more than 50% macroscopic fat [9], a thresh- old that is based on an Armed Forces Institute of Pathology case series [16]. This approach is a more cautious one than that suggested by the American College of Radiology, which suggests that the presence of macroscopic fat in an adrenal mass is a benign imaging find-
ing diagnostic of myelolipoma and indicates the need for no follow-up in incidentally dis- covered adrenal masses [17].
Nomenclature and definitions from the Society of Abdominal Radiology for fat in adrenal masses have been suggested [10]. The MRI appearance of macroscopic fat oc- curs because of the presence of a sufficient number of aggregated adipocytes and results in one or more of the following: intratumor- al signal intensity (SI) loss on imaging using fat-suppression techniques or chemical-shift artifact causing linear or curvilinear India ink (etching) artifact within or at the periph- ery of a mass at macroscopic fat-water inter- faces. Nonlinear, noncurvilinear SI loss on opposed-phase chemical-shift imaging (CSI) compared with in-phase CSI may be referred to as “microscopic fat” and is due to an in- sufficient amount of adipocytes or the pres- ence of fat within tumor cells [10].
It is not known if reports of macroscop- ic fat in ACC justify inclusion of ACC in the differential diagnosis of macroscopic fat-containing adrenal lesions or if a more cautious approach should be taken in ad- renal masses containing small amounts of macroscopic fat. Furthermore, a systematic evaluation of the validity of reports of ACC containing macroscopic fat has not been per- formed to our knowledge. The purpose of this study was to perform a systematic re- view to determine the number and quality of reports of ACC containing macroscopic fat; this information may inform guidelines for diagnosis and management of ACC.
Materials and Methods
This protocol was written using contemporary guidance regarding reporting and conduct of di- agnostic test accuracy imaging systematic re- views [18-24]. Reporting is done in accordance with Preferred Reporting Items for a Systematic Review and Meta-Analysis of Diagnostic Test Ac- curacy Studies (PRISMA-DTA) [21, 25]. Ethics approval is waived for this study type. The proto- col of this study was registered with PROSPERO (CRD42019119434) [26].
The following question was proposed using the mnemonic “PIT”: P (patients), patients with ACC; I (index test), MRI or CT of the adrenal glands; and T (target condition), the presence of macro- scopic fat in ACC confirmed at pathology. For the purpose of our study, ACC is defined as adrenal cortical neoplasms with at least one of the follow- ing: local invasion, distant metastases, or a Weiss score of 3 or greater [7, 8].
Search and Study Inclusion
A comprehensive search of PubMed, MEDLINE, Embase, and Cochrane databases was performed with the assistance of an experienced hospital librarian to identify studies evaluating the imaging appearance of ACC using the following terms: “adrenocortical cancer,” “adrenocortical carcinoma,” “adrenocortical neoplasm,” “adrenal cortex cancer,” “adrenal cortex carcinoma,” “ad- renal cortex neoplasm,” OR “ACC” in association with “tomography,” “CT,” “magnetic resonance,” OR “MRI.” The literature search included poten- tially relevant studies published from 1946 to No- vember 6, 2017.
All types of studies with full text in English language for evaluation of ACCs on CT or MRI in human subjects were included. The study had to describe macroscopic gross fat in at least one adrenal mass diagnosed as ACC on surgical pa- thology to be included. In the case of duplicate re- ports, the report with the most cases or the most detailed information allowing assessment of qual- ity was used.
Reviewer 1 (a 3rd-year diagnostic radiology resident) performed an initial title and abstract screen and discussed uncertainties with a fellow- ship-trained staff abdominal radiologist with 11 years of experience (reviewer 2). Non-ACC stud- ies, ACC studies without CT or MRI findings, du- plicate studies, and nonhuman studies were initially excluded. Studies without full text were excluded. The full text of the potentially relevant studies were retrieved, and inclusion criteria were independently applied by reviewer 1 and reviewer 3 (a 3rd-year di- agnostic radiology resident). Studies were excluded if any of the inclusion criteria were not met. Refer- ence lists of included studies were checked manu- ally to identify other relevant studies.
Data Extraction
From the included studies, the following infor- mation was extracted independently by the same reviewers who applied inclusion criteria using data extraction sheets (i.e., reviewers 1 and 3): name of the first author, journal, year of publication, coun- try of corresponding author’s institution, study de- sign, number of patients with ACC, patient age (mean and range), sex (proportions of male and female patients), lesion features (size, locally in- vasive features, metastatic disease), number of interpreting radiologists (reads done indepen- dently or in consensus and how were discrepan- cies resolved, years of experience, level of train- ing), and number of patients excluded (because of study overlap, insufficient test). In addition, the technical parameters of the imaging examinations were extracted. For MRI, reviewers extracted the
Macroscopic Fat in ACC
MEDLINE: 823 citations
Embase: 919 citations
PubMed: 13 citations
1274 Unique citations screened
The following inclusion and exclusion criteria were applied:
· Article must be an evaluation of ACC on CT or MRI in human subjects
771 Articles excluded after screening titles and abstracts
· Article must be written in English language
· Article may be a report of any study type
· Article must not be a duplicate of another study
503 Potentially relevant abstracts and 376 full-text articles retrieved in which macroscopic fat was reported in at least one adrenal mass diagnosed as ACC on the basis of surgical pathology results; inclusion and exclusion criteria were applied
499 Articles were excluded during data extraction:
· Full text was not available: 127 articles
· Full text was duplicated in another study: 2 articles
4 Articles included (3 Case reports with one case each and 1 retrospective study with four cases)
· Full text was not written in English language: 7 articles
· Final pathologic diagnosis was not ACC (e.g., adrenal oncocytoma): 33 articles
sequences used to diagnose macroscopic fat and the slice thickness; for CT, they extracted the slice thickness and whether contrast material was used. For all imaging examinations, reviewers record- ed the following information: the presence or ab- sence of macroscopic fat, definition for positivity used for macroscopic fat (size of region of fat, se- quences or method used to determine the presence of macroscopic fat), percentage of macroscopic fat in each lesion (if reported or estimated on the provided images), reference standard applied for confirmation of macroscopic fat and diagnosis of ACC, presenting symptoms, pathology report, and functionality of the tumor.
Piloting of the data extraction form was done with the first included study to ensure a high lev- el of agreement. Risk of bias was assessed using a modified Quality Assessment of Diagnostic Accu- racy Studies-2 (QUADAS-2) instrument tailored after pilot testing on the first two included stud- ies to evaluate three domains: flow and timing (i.e., Did all patients undergo the index test and refer- ence standard? Was there an appropriate interval between the index test and reference standard?), index test, and reference standard [27]. A low risk of bias for the index test domain is defined as un- equivocal presence of macroscopic fat documented on CT or MRI (i.e., definitive areas on CT at ← 20 HU or areas with loss of T1 signal on fat-saturation MRI sequences). A low risk of bias in the refer-
ence standard domain for diagnosis of ACC is an adrenocortical neoplasm at pathology showing at least one of the following: local invasion, distant metastases, or a Weiss score of 3 or greater. Any persistent disagreements between reviewers for study inclusion or data extraction after consensus discussion were resolved through discussion with reviewer 2.
Results
Search and application of the inclusion criteria yielded three case reports and one retrospective study (total of seven patients) of ACC with macroscopic fat on imaging (Fig. 1). Study characteristics are provided in Table 1. All seven patients presented with symptoms (lesion size, 6.5-22 cm); neither metastasis nor local invasive features were reported. In four patients, imaging (CT, n = 3; modality not reported, n = 1) showed less than 5% macroscopic fat. During data ex- traction, consensus of the extractors was reached for all cases except two, which re- quired discussion with reviewer 2.
We performed a risk-of-bias evaluation for the included studies. None of the included studies were at risk of bias in the flow and timing domain. Case 1 from Egbert et al. [12] had a high risk of bias for the index test and an unclear risk of bias for the reference stan-
dard domains. A potentially unreliable de- termination of macroscopic fat-a small ROI (0.34 cm2) and attenuation of -22 HU-was identified in this case [12]. This minute re- gion just below the threshold for diagnosis of macroscopic fat (typically -20 HU) rais- es the possibility that the finding is second- ary to image noise artifact rather than mac- roscopic fat. The reliability of the index test determination is further questioned because pathologic confirmation of macroscopic fat was not reported. Furthermore, detail re- garding pathologic diagnosis was insuffi- cient to determine whether our criteria for ACC diagnosis were met, which raises con- cern that case 1 could be an adenoma with myelolipomatous metaplasia.
Case 2 from Ferrozzi and Bova [13] was deemed at low risk of bias for the index do- main; a minute focus of macroscopic fat (<5% on review of the provided image) with an at- tenuation of -50 and -65 HU was identified on CT. The risk of bias for the reference stan- dard domain was determined to be unclear because detail regarding pathologic diagnosis was insufficient to determine whether our cri- teria for ACC diagnosis were met, which rais- es the concern that case 2 could be an adeno- ma with myelolipomatous metaplasia.
Case 3 from Heye et al. [14] had a low risk of bias for the index test domain. A 1.5 x 0.7 cm focus of macroscopic fat (< 5%) that showed the characteristic increased SI on T1-weighted images (isointense to retroperi- toneal fat) and homogeneous loss of SI with fat-suppression techniques was observed on MRI. The risk of bias for the reference stan- dard was unclear. The authors provided the following excerpts from the pathology re- port: “Histologic examination of the speci- men showed an ACC without evidence of vascular invasion but with a focus of mature adipose tissue that measured 1.5 × 0.7 cm, no morphologic or immunohistochemical evi- dence for myelolipoma … Myeloperoxidase staining to detect foci of myeloid cells were negative.” Like the reports of cases 1 and 2, the report of case 3 provided insufficient de- tail regarding pathologic diagnosis to deter- mine whether our criteria for ACC diagnosis were met, which raises concern that case 3 could represent an adenoma with myelolipo- matous metaplasia.
Cases 4-7 from Zhang et al. [15] described four cases with a low risk of bias for the in- dex test domain: Foci of macroscopic fat (at- tenuation range, -41 to -81 HU) on CT were present in each lesion. There was an unclear
| First Author [Reference No.] (Case No.) | No. of Lesions | Age (y) | Sex | Tumor Size (cm) | Tumor Characteristics | Index Test (Index Test Findings) | Symptoms | Presence of Invasive Features or Metastases | Percentage of Fat Seen on Imagingª | Reference Standard (Report Wording) |
|---|---|---|---|---|---|---|---|---|---|---|
| Egbert [12] (1) | 1 | 41 | M | 22 | Heterogeneous, peripheral nodular enhancement, central necrosis | CT (-22 HU for 0.34-cm2 ROI) | Weight loss, nausea, vomiting, abdominal pain | NR | <5 | Surgery (low-grade ACC)b |
| Ferrozzi [13] (2) | 1 | 62 | F | 6.5 | Well-defined contours, heterogeneous | CT (small focus with attenuation of-50 and -65 HU, similar to fat) | 2 months of right flank pain | NR | <5 | Surgery (ACC with small island of fatty tissue) |
| Heye [14] (3) | 1 | 38 | M | 7 | Heterogeneous, peripheral enhancement in arterial and venous phases | MRI (loss of SI on fat-saturation images in small [1.5 x 0.7 cm] focus) | Hypertension, presyncope, nausea, paresthesias, headaches, severe tinnitus | NR | <5 | Surgery (ACC with a marginal focus of mature adipose tissue, without vascular invasion; no morphologic or immunohistochemical evidence of myelolipoma) |
| Zhang [15] (4-7) | 4 | NR | NR | NR | One lesion had a well-defined, thin capsule; characteristics of three lesions NR | CT (foci of mature fat with attenuation ranging from -41 to -81 HU) | NR | NR | < 5 in one lesion (based on CT image); NR in other three lesions | Surgery for two lesions (no myelolipoma or other collision tumors containing fat were identified in the two specimens with fat foci that underwent adrenalectomy); biopsy for two lesions but no further details or histopathologic evaluation provided (NR) |
Note-NR = notreported, ACC = adrenocortical carcinoma, SI = signal intensity. ªThe imaging modality was CT in three patients and was NR in one patient.
bFat was NR.
risk of bias for the reference standard domain. The risk of bias for the reference standard domain was determined to be unclear. Only two of the four lesions had surgical pa- thology to confirm the presence of fat, and two had biopsy. Like the reports of cases 1-3, the report of cases 4-7 pro- vided insufficient detail regarding pathologic diagnosis to determine whether our criteria for ACC diagnosis were met, which raises concern that cases 4-7 could be adeno- mas with myelolipomatous metaplasia.
Discussion
Our systematic review identified four reports [12-15] of seven ACCs with macroscopic fat identified on imag- ing. All patients presented with symptoms, and all lesions were more than 6 cm in diameter. When the percentage fat on imaging was reported, four of seven lesions had less than 5% macroscopic fat on imaging. Key details regard- ing pathologic determination of ACC were scant: Neither the presence of metastatic disease nor local invasion was reported, and the system used for determination of ACC rather than adenoma (e.g., Weiss score) and the results of this classification system were not reported in any of these cases. Thus, there is an unclear risk of bias pertaining to the reference standard applied in these cases. This lack of detail regarding diagnostic criteria makes it impossible to determine if these lesions were truly ACCs or if they could be adenomas with myelolipomatous metaplasia. Unfortunately, a lack of complete reporting is common in the imaging diagnostic accuracy literature [28, 29]. This lack of detail limits the ability of readers to determine the validity of study findings. Given that macroscopic fat in ACC is such a rare finding, it is disappointing that critical details were missing from these study reports.
Macroscopic fat in an adrenal mass is generally considered diagnostic of adrenal myelolipomas [9, 17]. It is also known that myelolipomas can contain varying amount of fat and spe- cifically the group termed “adrenal myelolipoma with associ- ated adrenal disease” could contain less fat. In a case series of myelolipoma, Rao et al. [16] reported that 75% of lesions had less than 10% fat, lesions more frequently had calcifica- tion (56% of lesions), and lesions were typically smaller (size range not specified). Thus, despite the case reports identified in our systematic review, it remains overwhelmingly likely that any lesion with gross fat, even a small amount, represents a myelolipoma. It is noteworthy that in our series, all lesions were large (> 6 cm); symptomatic; and, when reported, had less than 5% macroscopic fat on imaging.
There are several limitations to our study. First, despite extensive searching and attempts to retrieve study reports, including interlibrary loans, we were unable to retrieve all full text reports; however, based on abstract perusal, it is unlikely that any of these studies would have been included at the level of the full-text screen; we used a high sensitiv- ity and low specificity inclusion for moving studies on to full text screen to minimize the probability of missed cases. Furthermore, our search is somewhat dated; however, the resources required to extensively screen the large number of potential studies required considerable time. Because
Macroscopic Fat in ACC
this area of research is not a rapidly evolving field, it is unlikely that the search dates missed a high number of additional potential reports. The level of detail provided in the included study reports was insufficient to determine key features regarding diagnosis because of a lack of complete reporting of the imaging diagnostic accuracy studies and the details of diagnostic criteria of the included cases.
In conclusion, there are few reports of macroscopic fat in ACC detected on imag- ing studies; tumors were large (> 6 cm) and had a small proportion of gross fat (< 5%). The reliability of reported cases is question- able because of the unclear or high risk of bias, which is largely due to lack of suffi- cient detail regarding pathologic diagnosis in the study reports. Based on this informa- tion, a change in guideline recommendations may not be warranted. However, until large case series with appropriate detail regarding pathologic diagnosis are performed, consid- eration of follow-up imaging or biopsy of pa- tients with large (> 6 cm) symptomatic tu- mors containing a small proportion of fat (< 5%) may be appropriate.
References
1. Jain M, Kapoor S, Mishra A, Gupta S, Agarwal A. Weiss criteria in large adrenocortical tumors: a validation study. Indian J Pathol Microbiol 2010; 53:222-226
2. Kebebew E, Reiff E, Duh QY, Clark OH, McMillan A. Extent of disease at presentation and outcome for adrenocortical carcinoma: have we made progress? World J Surg 2006; 30:872-878
3. Libé R. Adrenocortical carcinoma (ACC): diag- nosis, prognosis, and treatment. Front Cell Dev Biol 2015; 3:45
4. Else T, Rodriguez-Galindo C. Hereditary predis- position to childhood ACC and the associated mo- lecular phenotype: 5th International ACC Sympo- sium Session-not just for kids! Horm Cancer 2016; 7:36-39
5. Libé R. Adrenocortical carcinoma (ACC): diag- nosis, prognosis, and treatment. Front Cell Dev Biol 2015; 3:45
6. Fay AP, Elfiky A, Teló GH, et al. Adrenocortical carcinoma: the management of metastatic disease. Crit Rev Oncol Hematol 2014; 92:123-132
7. Aubert S, Wacrenier A, Leroy X, et al. Weiss sys- tem revisited: a clinicopathologic and immuno- histochemical study of 49 adrenocortical tumors. Am J Surg Pathol 2002; 26:1612-1619
8. Weiss LM. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocorti- cal tumors. Am J Surg Pathol 1984; 8:163-169
9. Schieda N, Siegelman ES. Update on CT and MRI of adrenal nodules. AJR 2017; 208:1206-1217
10. Schieda N, Davenport MS, Pedrosa I, et al. Renal and adrenal masses containing fat at MRI: pro- posed nomenclature by the Society of Abdominal Radiology disease-focused panel on renal cell car- cinoma. J Magn Reson Imaging 2019; 49:917-926
11. Schieda N, Al Dandan O, Kielar AZ, Flood TA, McInnes MD, Siegelman ES. Pitfalls of adrenal imaging with chemical shift MRI. Clin Radiol 2014; 69:1186-1197
12. Egbert N, Elsayes KM, Azar S, Caoili EM. Com- puted tomography of adrenocortical carcinoma containing macroscopic fat. Cancer Imaging 2010; 10:198-200
13. Ferrozzi F, Bova D. CT and MR demonstration of fat within an adrenal cortical carcinoma. Abdom Imaging 1995; 20:272-274
14. Heye S, Woestenborghs H, Van Kerkhove F, Oyen R. Adrenocortical carcinoma with fat inclusion: case report. Abdom Imaging 2005; 30:641-643
15. Zhang HM, Perrier ND, Grubbs EG, et al. CT fea- tures and quantification of the characteristics of ad- renocortical carcinomas on unenhanced and con- trast-enhanced studies. Clin Radiol 2012; 67:38-46
16. Rao P, Kenney PJ, Wagner BJ, Davidson AJ. Im- aging and pathologic features of myelolipoma. RadioGraphics 1997; 17:1373-1385
17. Mayo-Smith WW, Song JH, Boland GL, et al. Management of incidental adrenal masses: a white paper of the ACR Incidental Findings Com- mittee. J Am Coll Radiol 2017; 14:1038-1044
18. McGrath TA, Bossuyt PM, Cronin P, et al. Best practices for MRI systematic reviews and meta- analyses. J Magn Reson Imaging 2019; 49:e51-e64
19. McInnes MD, Bossuyt PM. Pitfalls of systematic reviews and meta-analyses in imaging research.
Radiology 2015; 277:13-21
20. Macaskill P, Gatsonis C, Deeks JJ, Harbord RM, Takwoingi Y. Analysing and presenting results. In: Deeks JJ, Bossuyt PM, Gatsonis C, eds. Co- chrane handbook for systematic reviews of diag- nostic test accuracy, version 1.0. London, UK: The Cochrane Collaboration, 2010
21. McGrath TA, Alabousi M, Skidmore B, et al. Rec- ommendations for reporting of systematic reviews and meta-analyses of diagnostic test accuracy: a systematic review. Syst Rev 2017; 6:194
22. McGrath TA, McInnes MD, Korevaar DA, Bossuyt PM. Meta-analyses of diagnostic accura- cy in imaging journals: analysis of pooling tech- niques and their effect on summary estimates of diagnostic accuracy. Radiology 2016; 281:78-85
23. McGrath TA, McInnes MD, Langer FW, Hong J, Korevaar DA, Bossuyt PM. Treatment of multiple test readers in diagnostic accuracy systematic re- views-meta-analyses of imaging studies. Eur J Radiol 2017; 93:59-64
24. McGrath TA, McInnes MD, van Es N, Leeflang MM, Korevaar DA, Bossuyt PM. Overinterpreta- tion of research findings: evidence of “spin” in systematic reviews of diagnostic accuracy studies. Clin Chem 2017; 63:1353-1362
25. Frank RA, Bossuyt PM, McInnes MD. Systematic reviews and meta-analyses of diagnostic test ac- curacy: the PRISMA-DTA Statement. Radiology 2018; 289:313-314
26. National Institute for Health Research website. PROSPERO: international prospective register of systematic reviews. www.crd.york.ac.uk/PROSPE- RO/. Published 2014. Accessed November 14, 2014
27. Whiting PF, Rutjes AW, Westwood ME, et al .; QUADAS-2 Group. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011; 155:529-536
28. Hong PJ, Korevaar DA, McGrath TA, et al. Re- porting of imaging diagnostic accuracy studies with focus on MRI subgroup: adherence to STARD 2015. J Magn Reson Imaging 2018; 47:523-544
29. Salameh JP, McInnes MD, Moher D, et al. Com- pleteness of reporting of systematic reviews of di- agnostic test accuracy based on the PRISMA-DTA reporting guideline. Clin Chem 2019; 65:291-301