Multi-Institutional Review of the Preoperative Diagnostic Accuracy for Pediatric Ovarian Mature Cystic Teratomas

Maria E. Knaus, Amanda J. Onwuka, Amin Afrazi, Peter C. Minneci, Katherine J. Deans

JPAG 2022; VOLUME 35, ISSUE 4, P478-485, AUGUST 2022

1. What was the primary objective of this study? What are the strengths and limitations of the studies methodological approach towards answering this objective? What do you think a stronger approach would be?

A: Primary objective is evaluating the preoperative diagnostic accuracy of current diagnostic techniques for MCT. Their approach is a multisite retrospective chart review. The strength of multiple sites is to reduce the bias of single site diagnostic approaches, however a limitation is the induction of greater heterogeneity with regards to clinical approaches and inconsistent reporting. A retrospective review provides existing outcomes which limits our ability to control how much diagnostic approach influenced the outcome. Additionally, the preoperative diagnosis could have been obtained from the operative report, which could then be influenced given the operative report is typically done after the procedure is done and when an answer is known.

A possible stronger study would be to randomize ultrasound suspected MCT to either be additionally discussed in a multi-disciplinary conference with tumor markers or not and evaluate outcomes. A prospective study would also help reduce biases and individual variations in management.

2. What imaging findings were most concerning for malignant masses?

A: Free fluid, solid components, papillary projections, irregular borders, extension into surrounding structures, lymphadenopathy, metastases, and larger size were more concerning for malignancy.

3. Laboratory testing can differ across institutions and even within a lab. Describe common tumor markers obtained for complex masses. Describe why laboratory testing may differ across institutions. How might alterations in the normal ranges of tumor markers affect the management of ovarian masses?

A. Tumor markers obtained in this study included AFP, CA 125, CA 19-9, CEA, Inhibin A, Inhibin B, LDH, and b-HCG. Laboratory testing can be influenced by the type of test that is being performed (e.g. immunoassay vs mass spectrometry). Additionally normal ranges may be defined by a specific patient population (including a hospital population, a national labs normative sample) and thus can differ depending on where a lab is getting its "normal" population. This can mean that in two separate labs they might have different cutoffs for their normal ranges of a specific tumor marker, leading a patient to appear elevated in one and normal range in another. This could lead a surgeon to classifying one patient as higher risk for a malignancy, even if they both have the same absolute value. Serum levels of tumor markers may also be affected by individual factors, including pubertal status, menstrual timing, and presence of inflammation for unrelated reasons.

4.Table 3 and 4 both describe univariate analysis. What are the limitations of using this approach in a study like this?

A. Univariate analyses compare a single variable between two or more groups. However, this does not take into account possible confounders that may affect any potential trends identified. In surgical decision making, early data may influence the choice of whether or not to acquire additional data (such as additional imaging, the use of consultation, or acquisition of tumor markers). It is unclear in any one measure if the others influenced a surgeon’s decision to acquire these additional measures thus providing bias for the outcome. Controlling for sequential acquisition of diagnostic tools could further help to identify the accuracy of a diagnostic tool in the context of how it was used. Additionally, as authors described, it's unclear how surgeons used these diagnostic techniques to inform their surgical preoperative decision making.

5. 4% of those who were diagnosed with MCT had malignant pathology. What do you think about that percentage? What are the risks of misdiagnosing malignant vs benign lesions?

A. Absolute percentages in clinical care must be taken in context. 4% of individuals with abnormal menses being misdiagnosed with PCOS is very different than a missed diagnosis of malignant pathology. The risk of misdiagnosis malignant lesions includes spread of pathology, possible loss of additional reproductive function with advanced surgical intervention and adjuvant therapy requirements that could have been avoided at earlier staging. The risk of misdiagnosis of benign lesions includes unnecessary oophorectomy or other risk reducing surgery which may include additional possible complications as well as implications on fertility. Journal club participations should consider what they believe to be "Acceptable risk" in this context.