MRI after Whoops procedure: diagnostic value for residual sarcoma and predictive value for an incomplete second resection

Abstract

Objective To determine the value of MRI for the detection and assessment of the anatomic extent of residual sarcoma after a Whoops procedure (unplanned sarcoma resection) and its utility for the prediction of an incomplete second resection. Materials and methods This study included consecutive patients who underwent a Whoops procedure, successively followed by gadolinium chelate-enhanced MRI and second surgery at a tertiary care sarcoma center. Results Twenty-six patients were included, of whom 19 with residual tumor at the second surgery and 8 with an incomplete second resection (R1: n = 6 and R2: n = 2). Interobserver agreement for residual tumor at MRI after a Whoops procedure was perfect (κ value: 1.000). MRI achieved a sensitivity of 47.4% (9/19), a specificity of 100% (7/7), a positive predictive value of 100% (9/9), and a negative predictive value of 70.0% (7/17) for the detection of residual tumor. MRI correctly classified 2 of 19 residual sarcomas as deep-seated (i.e., extending beyond the superficial muscle fascia) but failed to correctly classify 3 of 19 residual sarcomas as deep-seated. There were no significant associations between MRI findings (presence of residual tumor, maximum tumor diameter, anatomic tumor extent, tumor margins, tumor spiculae, and tumor tail on the superficial fascia) with an incomplete (R1 or R2) second resection. Conclusion Gadolinium chelate-enhanced MRI is a reproducible method to rule in residual sarcoma, but it is insufficiently accurate to rule out and assess the anatomic extent or residual sarcoma after a Whoops procedure. Furthermore, MRI has no utility in predicting an incomplete second resection.

Diagnostic performance of MRI in detecting residual soft tissue sarcoma after unplanned excision: Systematic review and meta-analysis

“…There was patient spectrum bias in four studies, because these studies also included some patients who underwent neoadjuvant therapy before re-excision [13,14] and some patients with other soft tissue malignancies than sarcoma [4,17]. In addition, in 4 of the included studies [12,14,16,17], MRI was reviewed by consensus of two interpreters, which does not mirror the independent readings that radiologists typically perform in clinical practice. Interobserver agreement for residual tumor at MRI was reported by only one study and was reported to be perfect (κ value of 1.00) [2].…”

“…This remains a topic for future study. Specificity of MRI was fairly high but not perfect, which may be explained by the presence of postoperative hematoma, granulation tissue or scarring which could in some cases be mistaken for residual soft tissue sarcoma [12,17,16,24]. Specificity may be improved by using image subtraction in gadolinium-enhanced MRI [12], diffusionweighted (DW) MRI, and/or dynamic contrast-enhanced (DCE) MRI [16,24].…”

“…After reading the full text version of the 11 remaining studies, one study was excluded because the diagnostic performance of MRI in detecting residual soft tissue sarcoma after unplanned excision was not investigated [11]. Eventually, 10 studies remained and were included [4,[12][13][14][15][16][17][18][19][20] (Fig. 1).…”

“…2. Risk of bias with respect to patient selection was rated low in all ten included studies [4,[12][13][14][15][16][17][18][19][20]. Risk of bias with respect to index test was rated unclear in five studies [4,13,15,16,19], because these studies did not report whether interpretation of MRI was blinded to the reference standard.…”

“…Similarly, risk of bias with respect to reference standard was rated unclear in all ten included studies [4,[12][13][14][15][16][17][18][19][20], because none reported whether interpretation of the reference standard was blinded to MRI findings. Risk of bias with respect to flow and timing was rated high in one study [16], because not all included patients received the same reference standard (pathologic findings after re-excision served as reference standard in 21 patients, whereas clinical/imaging follow-up served as reference standard in the remaining 17 patients).…”