Malignancy Assessment Using Gene Identification in Captured Cells Algorithm for the Prediction of Malignancy in Women With a Pelvic Mass : Obstetrics & Gynecology

Secondary Logo

Journal Logo

Contents: Original Research

Malignancy Assessment Using Gene Identification in Captured Cells Algorithm for the Prediction of Malignancy in Women With a Pelvic Mass

Moore, Richard George MD; Khazan, Negar PhD; Coulter, Madeline Ann MD; Singh, Rakesh PhD; Miller, Michael Craig BS; Sivagnanalingam, Umayal BS, MS; DuBeshter, Brent MD; Angel, Cynthia MD; Liu, Cici MD; Seto, Kelly PhD; Englert, David PhD; Meachem, Philip MS; Kim, Kyu Kwang PhD

Author Information
doi: 10.1097/AOG.0000000000004927
  • Open

The prevalence of pelvic masses is relatively common, with 5–10% of women being diagnosed with a mass at some point in their lifetime.1 Annually in the United States, more than 200,000 women undergo surgery for a pelvic mass, with 13–21% being malignant.2 It is critical to accurately assess the risk for malignancy in women who present with a pelvic mass before surgery, because morbidity and mortality outcomes are significantly improved when women with ovarian cancer have surgery performed by gynecologic oncologists experienced in the management of ovarian cancer.3–5 Currently, history and physical, imaging, serum CA 125 levels and formal biomarker algorithms are included in the pelvic mass assessment guidelines from the American College of Obstetricians and Gynecologists.6 Despite these guidelines, only 30–50% of women diagnosed with ovarian cancer are referred to gynecologic oncologists.7,8 Although improvements have been made in the triage of women presenting with a pelvic mass by the addition of U.S. Food and Drug Administration–cleared predictive algorithms such as ROMA® (Risk of Ovarian Malignancy Algorithm) and OVA1®, further improvements are needed.6,9–12

Emerging technologies for the isolation and interrogation of cells captured from blood, comprised of circulating tumor cells (CTCs) and other rare circulating cells (RCCs), present a novel opportunity for the detection and potential characterization of malignancy with a simple peripheral blood test.13 This study was designed to evaluate multiplexed gene expression of cancer related targets in cells captured from peripheral blood (CTCs and RCCs) using a liquid biopsy system, alone or in combination with serum biomarkers, for detection of malignancy in women with a pelvic mass.

METHODS

This prospective clinical study, registered with clinicaltrials.gov (NCT02781272), was approved by the University of Rochester Wilmot Cancer Institute IRB and conducted through the Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Wilmot Cancer Institute at the University of Rochester. Any patient diagnosed with an ovarian cyst or pelvic mass who was scheduled for an imaging guided biopsy, surgical biopsy, or surgical excision for definitive tissue diagnosis of their pelvic mass was eligible for enrollment into the trial. All patients were required to have a pelvic mass documented by imaging (ultrasonography, computed tomography, or magnetic resonance imaging) within 90 days before surgery or biopsy. Patients with a malignancy diagnosed within the prior 5 years were excluded from enrollment into the study, except for patients with skin cancers (squamous cell or basal cell).

A convenience sample of 200 patients with a pelvic mass was selected based on patient availability and the expected incidence of malignancy in this referred population of 20–30%. The goal was to obtain samples from a minimum of 50 women diagnosed with a malignant pelvic mass to evaluate the ability of using multiplexed gene expression evaluation of cells captured by the Parsortix® system from preoperatively collected peripheral blood (CTCs and RCCs) for the identification of malignancy in women with a pelvic mass.

Imaging guided biopsy, surgical biopsy, or surgical excision for evaluation of the pelvic mass was performed by a qualified physician. All tissue samples were evaluated by the gynecologic oncology pathology team at the University of Rochester pathology department. Results of the histopathologic evaluation, including the final diagnosis, along with histologic subtype and stage of cancer (if present), were documented. Only patients with a confirmed histologic diagnosis were considered evaluable. Race, age, and menopausal status information was recorded for all patients to determine the demographics of the population of women included in the study. All data collected were independently source verified.

Blood samples were obtained within 30 days before or on the day of surgery or biopsy of the pelvic mass and before anesthesia. Up to 35 mL of whole blood was collected by venipuncture into the following evacuated blood collection tubes for all patients: one 5-mL serum separating tube followed by three separate 10-mL K2EDTA tubes. The 5-mL serum separating tube was drawn first and was used for serum biomarker testing, followed by the collection of blood into the three K2EDTA tubes, which were pooled and used for the capture and harvest of circulating cells. The collection of the serum separating tube first helps to avoid contamination of the K2EDTA samples with squamous or noncirculating cells that might be collected as a result of the venipuncture. The blood collected into the K2EDTA tubes was pooled and two equal volume aliquots (7.5–15 mL each, 12.2 mL on average) were processed using Parsortix PR1 Systems for the capture of circulating cells (CTCs and RCCs) based on their size and deformability.14 The captured cells from each aliquot of blood were harvested into a single tube (each cell harvest consisted of 210 microliters of phosphate-buffered saline, so a combined harvest contained 420 microliters of phosphate-buffered saline and cells), and the captured circulating cells were lysed with RLT buffer containing 1% 2-mercaptoethanol (if a single harvest was collected, 600 microliters of RLT buffer was added to the harvest [total lysate volume of 810 microliters]; if two harvests were combined, 1,200 microliters of RLT buffer was added to the combined harvest [total lysate volume of 1,410 microliters]). The cell lysate from each patient was split into two equal volume aliquots and stored frozen at −180°C in liquid nitrogen. Using the RNeasy Micro Kit, RNA was purified from one of the lysate aliquots and subsequently analyzed using a highly multiplexed gene expression assay to evaluate the expression of 72 different gene transcripts representing 52 different ovarian cancer–associated genes and eight different housekeeping genes (Table 1). The ovarian cancer–associated genes were identified through a literature search as being primarily related to ovarian cancer or to cancer in general. The housekeeping genes were included to help gauge the quality of the RNA obtained, as well as the relative quantity of blood cells (primarily white blood cells) present in the lysates.

T1
Table 1.:
Gene Transcripts Analyzed

Serum obtained from blood collected into the 5-mL serum separating tube was used to measure seven different serum biomarkers related to ovarian cancer and menopausal status that are currently in use in one or more multivariate biomarker algorithms for pelvic mass risk assessment. Serum levels of CA 125, HE4, transthyretin (prealbumin), apolipoprotein A1, transferrin, β-2-microglobulin, and follicle-stimulating hormone were determined for each patient. All serum biomarker testing was performed through the University of Rochester Laboratories and ARUP Laboratories.

The primary objective of the study was to evaluate the ability of cancer-related gene expression in circulating cells captured from peripheral blood using a liquid biopsy system, alone or in combination with serum biomarkers, to detect malignancy in patients with a pelvic mass. Univariate logistic regression and multivariate backward-stepwise logistic regression analyses of the gene expression results and the serum biomarkers, alone and in combination, were performed to develop algorithms for the differentiation of various histopathologic diagnosis groupings of the evaluable patients (eg, benign or low malignant potential vs all cancers, benign vs all cancers). Receiver operating characteristic curves were constructed for each univariate and multivariate model, the area under the curve (ROC-AUC) was determined, and the equality of the ROC-AUC was tested (“roccomp” command in Stata). Sensitivity, specificity, positive predictive value, negative predictive value, likelihood ratio of a positive test, likelihood ratio of a negative test, and overall accuracy for the multivariate logistic regression models were determined using various thresholds for the predictive probabilities.

RESULTS

A total of 200 patients with an ovarian cyst or pelvic mass were enrolled into the study between June 2016 and April 2017, of which 183 (91.5%) were evaluable. The mean and median age of the evaluable cohort was 56 years (range: 19–91). Of these patients, 68 (37.2%) were premenopausal and 115 (62.8%) were postmenopausal as determined through history and clinical exam. Demographics and final pathology diagnoses for the 183 evaluable patients are summarized in Tables 2 and 3.

T2
Table 2.:
Patient Demographics
T3
Table 3.:
Summary of Histologic Subtypes for Benign and Malignant Tumors

Of the 183 evaluable patients, 104 (56.8%) were diagnosed with benign disease, 17 (9.3%) were diagnosed with low malignant potential tumors, 42 (23.0%) were diagnosed with ovarian cancer, 14 (7.6%) were diagnosed with nonovarian gynecologic cancer, and six (3.3%) were diagnosed with nongynecologic metastatic cancers. For the 37 (20.1%) patients diagnosed with invasive epithelial ovarian cancer, 12 (32.4%) had a stage I disease, four (10.8%) had stage II, nine (24.4%) had stage III, and 12 (32.4%) had stage IV.

The areas under the receiver operating characteristic curve (ROC-AUC) of the individual genes and serum biomarkers for their ability to discriminate between various histopathologic diagnosis groupings of the 183 evaluable patients are reported in Table 4. Analysis of the expression of individual genes in the captured circulating cells (CTCs and RCCs) through RNA interrogation for the prediction of benign (excluding low malignant potential tumors) compared with malignancy (all cancers) did not produce any significant results when examining each gene alone. The highest ROC-AUC found for any one individual gene was seen for FN1 (transcript 1) when differentiating between benign or low malignant potential compared with all cancers, with an ROC-AUC of 65.2% (95% CI 56.6–73.7%), followed by CCR2 (transcript 2) and CD274 when differentiating between benign compared with all cancers, with ROC-AUCs of 64.4% (95% CI 55.9–72.9%) and 64.2% (95% CI 55.6–72.9%), respectively. The individual serum biomarkers had higher ROC-AUCs than any of the individual genes, with the highest being HE4 in all cases, but particularly when differentiating between benign or low malignant potential compared with epithelial ovarian cancer only with an ROC-AUC of 89.4% (95% CI 83.4–95.3%).

T4
Table 4.:
Area Under the Receiver Operating Characteristic Curve Analysis of the Individual Genes and Serum Biomarkers for Their Ability to Differentiate Between Various Histopathologic Diagnosis Groupings of the Evaluable Patients

Backward-stepwise multivariate logistic regression analysis of the gene expression results alone, the serum biomarkers alone, and the gene expression results and serum biomarkers in combination for their ability to discriminate between various histopathologic diagnosis groupings of the 183 evaluable patients are reported in Table 5. Comparison of the ROC-AUCs between the gene-only and serum biomarkers–only multivariate algorithms for the discrimination between the various histopathologic diagnosis groupings showed nonsignificant differences (all P>.38). However, combining gene expression analysis with serum biomarker analysis produced the highest ROC-AUCs compared with either serum markers alone or gene expression results alone (Table 5). In most instances, a comparison of the ROC-AUCs for the gene-only and serum biomarkers–only algorithms to the combination algorithms showed significant differences (P<.05). The combination of genes and serum biomarkers to differentiate benign compared with all cancers showed that a multivariate model combining the expression of eight genes and four serum protein biomarkers achieved the highest ROC-AUC. This algorithm, referred to hereafter as the MAGIC (Malignancy Assessment using Gene Identification in Captured Cells) algorithm, was composed of the gene expression of PPIA, TBP, TPT1 (transcripts 2 and 3), WFDC-2 (HE4), INHA, VEGFA, CCR2, and SEPT2, plus serum levels of β-2-microglobulin (B2M), transferrin, CA 125 and HE4, and achieved an ROC-AUC of 95.1% (95% CI 92.0–98.2%) for the discrimination of benign tumors from all cancers. The MAGIC algorithm significantly outperformed all individual genes (ROC-AUC 50.2–65.2%; all P<.001), the serum biomarkers–only algorithm (ROC-AUC 89.6%, 95% CI 84.3–95.0%), and the gene-only algorithm (ROC-AUC 88.0, 95% CI 82.9–93.0%, P=.005) for the discrimination of benign tumors from all cancers (Table 5 and Fig. 1).

T5
Table 5.:
Area Under the Receiver Operating Characteristic Curve Analysis and Comparisons of the Multivariate Algorithms for the Gene Expression Results Alone, the Serum Biomarkers Alone, and the Combination of Gene Expression and Serum Biomarkers Results for Their Ability to Differentiate Between Various Histopathologic Diagnosis Groupings of the Evaluable Patients
F1
Fig. 1.:
Comparison of areas under receiver operating characteristic curves for the genes-only algorithm (green), the serum biomarkers–only algorithm (red), and the MAGIC (Malignancy Assessment Using Gene Identification in Captured Cells) algorithm (genes and serum biomarkers) (orange). Benign (n=104) vs all cancers (n=62).

Table 6 provides the estimates of sensitivity, specificity, negative predictive value, positive predictive value, accuracy, likelihood ratio of a positive test, and likelihood ratio of a negative test for the MAGIC algorithm, the serum biomarkers–only algorithm, and the gene expression–only algorithm at set specificities of approximately 65%, approximately 75%, approximately 85%, 95%, and 100% for the discrimination of benign tumors (excluding low malignant potential tumors) compared with any malignancy. Further, the MAGIC algorithm achieved an ROC-AUC of 89.5% (95% CI 81.3–97.8%) for patients with stage I–II epithelial ovarian cancer and 98.9% (95% CI 96.7–100%) for patients with stage III–IV disease (data not shown).

T6
Table 6.:
Performance Parameters of Sensitivity, Specificity, Positive Predictive Value, Negative Predictive Value, Accuracy, Likelihood Ratio of a Positive Test, and Likelihood Ratio of a Negative Test for the Following Multivariate Algorithms for the Differentiation of Benign Compared With All Cancers: the MAGIC (Malignancy Assessment Using Gene Identification in Captured Cells) Algorithm, the Serum Biomarkers–Only Algorithm, and the Gene Expression–Only Algorithm

DISCUSSION

Accurate risk stratification for patients presenting with a pelvic mass is essential for optimum management. Surgical intervention for patients with ovarian cancer performed by a gynecologic oncologist is associated with improved morbidity and mortality and better overall survival.5,7,15–20 Despite this, it is estimated that only 30–40% of patients with an ovarian malignancy undergo their initial surgery by a gynecologic oncologist.7,20–23 In recent years, there has been significant research in the development multiple biomarker algorithms to predict malignancy in patients with a pelvic mass. The serum biomarkers HE4 and CA 125 havebeen shown to be the markers with the most value for detecting malignancies when used in multiple marker algorithms.12,24–26 Analysis of serum biomarkers measured in the current trial found similar results to previous studies, which showed that the addition of other biomarkers to HE4 and CA 125 do not significantly improve the performance over the dual marker combination of HE4 and CA 125 for the differentiation of benign masses from malignant pelvic masses.25–30 The additive properties of serum biomarkers to detect malignancy may open the door to markers that are associated with, but perhaps not specific to, epithelial ovarian cancer. Although this could contribute to overall improvement in sensitivity of an algorithm, it could also lead to decreased specificity and little incremental value to the statistical performance of a dual marker algorithm.25 For these reason,s novel methods for detecting malignancies outside of serum biomarker analysis is needed.

Circulating tumor cells are cells that originate and detach from solid tumors and can be detected in the peripheral circulation of patients with malignancies. Circulating tumor cells are rare and have been estimated to have a ratio of 1 in 100 million to 1 in 1 billion cells in circulating blood.31,32 Multiple techniques and newer technologies have been developed to capture CTCs, which have now been demonstrated to be present in the blood of nearly all late-stage solid tumor patients and have also been shown to be present in early-stage disease.33–35 The presence of CTCs in blood has been shown to be a useful prognostic indicator in patients with primary and metastatic cancers.36–38 Gene expression profiling has been used in prognostic calculators in malignancies, such as multiple myeloma, breast cancer, and gliomas.39–42 The Parsortix system used in the current study employs a microfluidic based particle separation technology, with a gradient stepwise cassette that captures cells based on their size and deformability. Most blood cells pass across the steps and through the terminal gap and cells with increased size and rigidity are captured on the cassette and preserved. This system has been shown to efficiently capture RCCs and CTCs that can then be harvested and used for gene expression analysis.14,43 The liquid-biopsy and multiplex gene expression technologies used in the current trial are being investigated for detection of multiple cancer types, including ovary, endometrial, breast, prostate, and lung cancer.

In ovarian cancer, gene expression in CTCs is an emerging field not yet incorporated into existing clinical risk-stratification models. Recently, research has identified that in addition to transcoelomic spread, there is significant hematogenous spread, which may play a significant role in the metastatic spread to the omentum.44,45 Circulating tumor cell isolation, enrichment, and molecular characterization has been reported in patients with ovarian cancer. Kolostova et al reported on 40 patients using a cytomorphologic approach to identify captured cells for gene expression analysis. Comparison of the relative gene expression level in peripheral blood samples confirmed a statistically significant difference for gene expression of KRT7, WT1, EPCAM, MUC16, MUC1, KRT18 and KRT19. The combination of these genes could suggest the presence of ovarian cancer CTCs in the peripheral blood of patients.46

The role of CTCs, and multiple CTC gene expression identification and quantification analyses in epithelial ovarian cancer have been explored. However, no studies have incorporated the CTC gene expression analysis and serum biomarker levels into a risk assessment tool such as the one constructed in this study. We evaluated the potential of combining the gene expression of circulating cells captured from blood by liquid biopsy (CTCs and RCCs) with serum biomarker levels for the detection of cancer in women with a pelvic mass before surgical intervention. Single genes and their predictive performances were analyzed. The highest performing individual gene was FN1 with an AUC of 65.2%, followed by CCR2 with an AUC of 64.4%. CCR2 is the receptor for haptoglobin, which has been used as a nonspecific marker for inflammation, infection, and malignancy, including ovarian cancer. CCR2 may have relation to the tumor microenvironment and cell migration, which could be why it performed so well compared with other single genes. However, its significance and clinical application in ovarian cancer are still unclear.47 Similarly, multivariate analysis of the gene expression results alone compared with the serum biomarkers alone revealed that no combination outperformed the dual marker combination of HE4 and CA 125 for the prediction of benign (excluding low malignant potential tumors) compared with malignancy (all cancers and ovarian cancers only).

However, the unique grouping of eight genes with a total of nine different transcripts (PPIA, TBP, TPT1 transcripts 2 and 3, WFDC-2 [HE4], INHA, VEGFA, CCR2, and SEPT2) and four serum proteins (β-2-microglobulin, transferrin, CA 125, and HE4) comprising the MAGIC algorithm produced the highest ROC-AUC in this study (95.1%), which significantly outperformed any other gene or biomarker combination, including HE4 and CA 125. The MAGIC algorithm not only detected epithelial ovarian cancer, but also detected nonovary primary cancers, including metastatic cancers. At the same time, the MAGIC algorithm did not lose its performance regarding detection of early stage epithelial ovarian cancer, with an ROC-AUC for stage I–II disease of 89.5%. The inclusion of the WFDC-2 gene, which produces the protein HE4, is of interest as HE4 has recently been shown to play an inhibitory role of CD8+ infiltrating lymphocytes and inducing macrophages to the M2 nonactive form in the ovarian cancer tumor microenvironment.48 Approximately 80% of epithelial ovarian cancers overexpress the HE4 protein, which may speak to the importance of the WFCD-2 gene in pathogenesis.

A major strength of this study is that it is a large prospective clinical trial employing novel technology with the ability to capture circulating cells from blood (RCCs and CTCs) and analyze their gene expression. Often, biomarker and gene expression research are executed retrospectively, by using tissue biopsy samples through microarray analysis. All the patients in this study had gene expression testing from a prospective liquid biopsy of peripheral blood samples, in addition to serum testing using blood samples collected just before surgery. Moreover, the nature of biomarker and gene expression testing is inherently objective. The key measures for this model are not reliant on subjective factors such as physician assessment or imaging interpretation, which have been used by other risk-assessment models and, thus, subject to interpretive variation.49–51 Additionally, the MAGIC algorithm has the advantage that it can also detect cancers of nonovarian origin, including metastatic cancers. A primary limitation of this study is that it was performed at a single tertiary care institution with patients who had been referred to the division of gynecologic oncology, and, therefore, the cohort may not represent the general population. The MAGIC algorithm needs to be verified and validated in a separate real-world cohort comprised of patients from the general community with a pelvic mass of unclear risk of malignancy. Proof of efficacy in a prospective validation data set would be necessary before clinical application.

Authors' Data Sharing Statement

  • Will individual participant data be available (including data dictionaries)? No.
  • What data in particular will be shared? Not available.
  • What other documents will be available? Not available.
  • When will data be available (start and end dates)? Not applicable.
  • By what access criteria will data be shared (including with whom, for what types of analyses, and by what mechanism)? Not applicable.

REFERENCES

1. Curtin JP. Management of the adnexal mass. Gynecol Oncol 1994;55:S42–6. doi: 10.1006/gyno.1994.1340
2. Bast RC Jr, Skates S, Lokshin A, Moore RG. Differential diagnosis of a pelvic mass: improved algorithms and novel biomarkers. Int J Gynecol Cancer 2012;22(suppl 1):S5–8. doi: 10.1097/IGC.0b013e318251c97d
3. Bristow RE, Tomacruz RS, Armstrong DK, Trimble EL, Montz FJ. Survival effect of maximal cytoreductive surgery for advanced ovarian carcinoma during the platinum era: a meta-analysis. J Clin Oncol 2002;20:1248–59. doi: 10.1200/JCO.2002.20.5.1248
4. Eisenkop SM, Spirtos NM, Montag TW, Nalick RH, Wang HJ. The impact of subspecialty training on the management of advanced ovarian cancer. Gynecol Oncol 1992;47:203–9. doi: 10.1016/0090-8258(92)90107-t
5. Tingulstad S, Skjeldestad FE, Hagen B. The effect of centralization of primary surgery on survival in ovarian cancer patients. Obstet Gynecol 2003;102:499–505. doi: 10.1016/s0029-7844(03)00579-9
6. Evaluation and management of adnexal masses. Practice Bulletin No. 174. American College of Obstetricians and Gynecologists. Obstet Gynecol 2016;128:e210–26. doi: 10.1097/AOG.0000000000001768
7. Carney ME, Lancaster JM, Ford C, Tsodikov A, Wiggins CL. A population-based study of patterns of care for ovarian cancer: who is seen by a gynecologic oncologist and who is not? Gynecol Oncol 2002;84:36–42. doi: 10.1006/gyno.2001.6460
8. Gershenson DM. Why American women are not receiving state-of-the-art gynecologic cancer care. Cancer J 2001;7:450–7.
9. Bristow RE, Smith A, Zhang Z, Chan DW, Crutcher G, Fung ET, et al. Ovarian malignancy risk stratification of the adnexal mass using a multivariate index assay. Gynecol Oncol 2013;128:252–9. doi: 10.1016/j.ygyno.2012.11.022
10. Longoria TC, Ueland FR, Zhang Z, Chan DW, Smith A, Fung ET, et al. Clinical performance of a multivariate index assay for detecting early-stage ovarian cancer. Am J Obstet Gynecol 2014;210:78.e1–9. doi: 10.1016/j.ajog.2013.09.017
11. Moore RG, McMeekin DS, Brown AK, Disilvestro P, Miller MC, Allard WJ, et al. A novel multiple marker bioassay utilizing HE4 and CA125 for the prediction of ovarian cancer in patients with a pelvic mass. Gynecol Oncol 2009;112:40–6. doi: 10.1016/j.ygyno.2008.08.031
12. Moore RG, Miller MC, Disilvestro P, Landrum LM, Gajewski W, Ball JJ, et al. Evaluation of the diagnostic accuracy of the risk of ovarian malignancy algorithm in women with a pelvic mass. Obstet Gynecol 2011;118:280–8. doi: 10.1097/AOG.0b013e318224fce2
13. Lianidou E, Pantel K. Liquid biopsies. Genes Chromosomes Cancer 2019;58:219–32. doi: 10.1002/gcc.22695
14. Miller MC, Robinson PS, Wagner C, O'Shannessy DJ. The ParsortixTM Cell Separation System - a versatile liquid biopsy platform. Cytometry A 2018;93:1234–9. doi: 10.1002/cyto.a.23571
15. NIH consensus conference. Ovarian cancer. Screening, treatment, and follow-up. NIH Consensus Development Panel on Ovarian Cancer. JAMA 1995;273:491–7.
16. Vernooij F, Witteveen PO, Verweij E, van der Graaf Y, Heintz AP. The impact of hospital type on the efficacy of chemotherapy treatment in ovarian cancer patients. Gynecol Oncol 2009;115:343–8. doi: 10.1016/j.ygyno.2009.08.018
17. Goff BA, Matthews BJ, Larson EH, Andrilla CH, Wynn M, Lishner DM, et al. Predictors of comprehensive surgical treatment in patients with ovarian cancer. Cancer 2007;109:2031–42. doi: 10.1002/cncr.22604
18. The role of the obstetrician-gynecologist in the early detection of epithelial ovarian cancer in women at average risk. Committee Opinion No. 716. American College of Obstetricians and Gynecologists. Obstet Gynecol 2017;130:e146–9. doi: 10.1097/AOG.0000000000002299
19. Engelen MJ, Kos HE, Willemse PH, Aalders JG, de Vries EG, Schaapveld M, et al. Surgery by consultant gynecologic oncologists improves survival in patients with ovarian carcinoma. Cancer 2006;106:589–98. doi: 10.1002/cncr.21616
20. Earle CC, Schrag D, Neville BA, Yabroff KR, Topor M, Fahey A, et al. Effect of surgeon specialty on processes of care and outcomes for ovarian cancer patients. J Natl Cancer Inst 2006;98:172–80. doi: 10.1093/jnci/djj019
21. Chan JK, Kapp DS, Shin JY, Husain A, Teng NN, Berek JS, et al. Influence of the gynecologic oncologist on the survival of ovarian cancer patients. Obstet Gynecol 2007;109:1342–50. doi: 10.1097/01.AOG.0000265207.27755.28
22. Chao X, Wang X, Xiao Y, Ji M, Wang S, Shi H, et al. Clinical analysis of high risk factors for pelvic malignant tumors after hysterectomy for benign diseases. Medicine (Baltimore) 2019;98:e17540. doi: 10.1097/MD.0000000000017540
23. Rim SH, Hirsch S, Thomas CC, Brewster WR, Cooney D, Thompson TD, et al. Gynecologic oncologists involvement on ovarian cancer standard of care receipt and survival. World J Obstet Gynecol 2016;5:187–96. doi: 10.5317/wjog.v5.i2.187
24. Montagnana M, Danese E, Ruzzenente O, Bresciani V, Nuzzo T, Gelati M, et al. The ROMA (Risk of Ovarian Malignancy Algorithm) for estimating the risk of epithelial ovarian cancer in women presenting with pelvic mass: is it really useful? Clin Chem Lab Med 2011 49:521–5. doi: 10.1515/CCLM.2011.075
25. Moore RG, Blackman A, Miller MC, Robison K, DiSilvestro PA, Eklund EE, et al. Multiple biomarker algorithms to predict epithelial ovarian cancer in women with a pelvic mass: can additional makers improve performance? Gynecol Oncol 2019;154:150–5. doi: 10.1016/j.ygyno.2019.04.006
26. Moore RG, Brown AK, Miller MC, Skates S, Allard WJ, Verch T, et al. The use of multiple novel tumor biomarkers for the detection of ovarian carcinoma in patients with a pelvic mass. Gynecol Oncol 2008;108:402–8. doi: 10.1016/j.ygyno.2007.10.017
27. Grenache DG, Heichman KA, Werner TL, Vucetic Z. Clinical performance of two multi-marker blood tests for predicting malignancy in women with an adnexal mass. Clin Chim Acta 2015;438:358–63. doi: 10.1016/j.cca.2014.09.028
28. Leung F, Bernardini MQ, Brown MD, Zheng Y, Molina R, Bast RC Jr, et al. Validation of a novel biomarker panel for the detection of ovarian cancer. Cancer Epidemiol Biomarkers Prev 2016;25:1333–40. doi: 10.1158/1055-9965.Epi-15-1299
29. Nolen B, Velikokhatnaya L, Marrangoni A, De GK, Lomakin A, Bast RC Jr, et al. Serum biomarker panels for the discrimination of benign from malignant cases in patients with an adnexal mass. Gynecol Oncol 2010;117:440–5. doi: 10.1016/j.ygyno.2010.02.005
30. Shaw R, Lokshin AE, Miller MC, Messerlian-Lambert G, Moore RG. Stacking machine learning algorithms for biomarker-based preoperative diagnosis of a pelvic mass. Cancers (Basel) 2022;14:1291. doi: 10.3390/cancers14051291
31. Nelson NJ. Circulating tumor cells: will they be clinically useful? J Natl Cancer Inst 2010;102:146–8. doi: 10.1093/jnci/djq016
32. Yu M, Stott S, Toner M, Maheswaran S, Haber DA. Circulating tumor cells: approaches to isolation and characterization. J Cell Biol 2011;192:373–82. doi: 10.1083/jcb.201010021
33. Millner LM, Linder MW, Valdes R Jr. Circulating tumor cells: a review of present methods and the need to identify heterogeneous phenotypes. Ann Clin Lab Sci 2013;43:295–304.
34. Chemi F, Mohan S, Guevara T, Clipson A, Rothwell DG, Dive C. Early dissemination of circulating tumor cells: biological and clinical insights. Front Oncol 2021;11:672195. doi: 10.3389/fonc.2021.672195
35. Ried K, Eng P, Sali A. Screening for circulating tumour cells allows early detection of cancer and monitoring of treatment effectiveness: an observational study. Asian Pac J Cancer Prev 2017;18:2275–85. doi: 10.22034/APJCP.2017.18.8.2275
36. Ma X, Xiao Z, Li X, Wang F, Zhang J, Zhou R, et al. Prognostic role of circulating tumor cells and disseminated tumor cells in patients with prostate cancer: a systematic review and meta-analysis. Tumour Biol 2014;35:5551–60. doi: 10.1007/s13277-014-1731-5
37. Wang J, Wang K, Xu J, Huang J, Zhang T. Prognostic significance of circulating tumor cells in non-small-cell lung cancer patients: a meta-analysis. PLoS One 2013;8:e78070. doi: 10.1371/journal.pone.0078070
38. Zhang L, Riethdorf S, Wu G, Wang T, Yang K, Peng G, et al. Meta-analysis of the prognostic value of circulating tumor cells in breast cancer. Clin Cancer Res 2012;18:5701–10. doi: 10.1158/1078-0432.CCR-12-1587
39. Sgroi DC, Carney E, Zarrella E, Steffel L, Binns SN, Finkelstein DM, et al. Prediction of late disease recurrence and extended adjuvant letrozole benefit by the HOXB13/IL17BR biomarker. J Natl Cancer Inst 2013;105:1036–42. doi: 10.1093/jnci/djt146
40. Reme T, Hugnot JP, Bieche I, Rigau V, Burel-Vandenbos F, Prevot V, et al. A molecular predictor reassesses classification of human grade II/III gliomas. PLoS One 2013;8:e66574. doi: 10.1371/journal.pone.0066574
41. Vallon-Christersson J, Häkkinen J, Hegardt C, Saal LH, Larsson C, Ehinger A, et al. Cross comparison and prognostic assessment of breast cancer multigene signatures in a large population-based contemporary clinical series. Sci Rep 2019;9:12184. doi: 10.1038/s41598-019-48570-x
42. van Laar R, Flinchum R, Brown N, Ramsey J, Riccitelli S, Heuck C, et al. Translating a gene expression signature for multiple myeloma prognosis into a robust high-throughput assay for clinical use. BMC Med Genomics 2014;7:25. doi: 10.1186/1755-8794-7-25
43. Xu L, Mao X, Imrali A, Syed F, Mutsvangwa K, Berney D, et al. Optimization and evaluation of a novel size based circulating tumor cell isolation system. PLoS One 2015;10:e0138032. doi: 10.1371/journal.pone.0138032
44. Motohara T, Masuda K, Morotti M, Zheng Y, El-Sahhar S, Chong KY, et al. An evolving story of the metastatic voyage of ovarian cancer cells: cellular and molecular orchestration of the adipose-rich metastatic microenvironment. Oncogene 2019;38:2885–98. doi: 10.1038/s41388-018-0637-x
45. Pradeep S, Kim SW, Wu SY, Nishimura M, Chaluvally-Raghavan P, Miyake T, et al. Hematogenous metastasis of ovarian cancer: rethinking mode of spread. Cancer Cell 2014;26:77–91. doi: 10.1016/j.ccr.2014.05.002
46. Kolostova K, Pinkas M, Jakabova A, Pospisilova E, Svobodova P, Spicka J, et al. Molecular characterization of circulating tumor cells in ovarian cancer. Am J Cancer Res 2016;6:973–80.
47. Zhao C, Annamalai L, Guo C, Kothandaraman N, Koh SC, Zhang H, et al. Circulating haptoglobin is an independent prognostic factor in the sera of patients with epithelial ovarian cancer. Neoplasia 2007;9:1–7. doi: 10.1593/neo.06619
48. Rowswell-Turner RB, Singh RK, Urh A, Yano N, Kim KK, Khazan N, et al. HE4 overexpression by ovarian cancer promotes a suppressive tumor immune microenvironment and enhanced tumor and macrophage PD-L1 expression. J Immunol 2021;206:2478–88. doi: 10.4049/jimmunol.2000281
49. Bailey J, Tailor A, Naik R, Lopes A, Godfrey K, Hatem HM, et al. Risk of malignancy index for referral of ovarian cancer cases to a tertiary center: does it identify the correct cases? Int J Gynecol Cancer 2006;16(suppl 1):30–4. doi: 10.1111/j.1525-1438.2006.00468.x
50. Jacobs I, Oram D, Fairbanks J, Turner J, Frost C, Grudzinskas JG. A risk of malignancy index incorporating CA 125, ultrasound and menopausal status for the accurate preoperative diagnosis of ovarian cancer. Br J Obstet Gynaecol 1990;97:922–9. doi: 10.1111/j.1471-0528.1990.tb02448.x
51. Manjunath AP, Pratapkumar Sujatha K, Vani R. Comparison of three risk of malignancy indices in evaluation of pelvic masses. Gynecol Oncol 2001;81:225–9. doi: 10.1006/gyno.2001.6122

Supplemental Digital Content

Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.