Innovative Approaches to Gynecologic Cancer Prognosis Fueled by Metabolomics

Mass spectrometric technology has made headway with its ability to quantify and categorize metabolic signatures using machine learning algorithms.¹ This progress suggests earlier detection and predictability of survival in certain cancers; however, further research is needed to bring this technique into clinical practice. To understand this approach, Robert Nagourney, MD, and his team assessed metabolic signatures in gynecologic malignancies to predict resistance to standard-of-care platinum therapy.²

Beginning with gynecologic cancers, this 3-part series will explore early advancements in metabolomics for gynecologic, breast, and pancreatic cancers and what they could mean for future oncology treatment.

“[This method] opens the opportunity to quantify biology at a metabolic level, and I think that is the breakthrough. We now have algorithms [that] in some circumstances are perfectly accurate, [with] no errors in identifying the presence of cancer and, more importantly, prognosis. This opens interesting opportunities,” Nagourney said.

The Study

In a small study of 47 patients, Nagourney and team evaluated blood and tissue samples from those with adenocarcinoma of the ovary or uterus who were candidates for carboplatin plus paclitaxel treatment.² Investigators identified patients at the highest risk of relapse and death using metabolic signatures to predict platinum resistance.

Of these patients, 27 (64.3%) achieved complete remission following treatment. The mean time to disease progression was 1.9 years, with disease-free survival at 1.7 years and overall survival at 2.6 years.

Of particular interest were the findings related to cisplatin. The mean lethal concentration of 50% (LC50 ) of cisplatin was determined to be 1.15 μg/mL, with values ranging from 0.4 to 3.1 μg/mL. The investigators observed correlations between cisplatin LC₅₀ levels and treatment outcomes. Although not statistically significant, there was a trend toward decreased complete remission rates associated with higher cisplatin LC₅₀ values, and relative risk was equal to 0.76 (95% CI, 0.46-1.27).

Patients with higher cisplatin LC₅₀ levels experienced shorter disease-free survival times, indicating a potential association between drug resistance and treatment response. Further analysis showed connections between cisplatin LC₅₀ values and 186 metabolic signatures. The study data identified specific metabolic signatures, such as lipid ratios and amino acid levels, that demonstrated predictive ability regarding the risk of relapse and mortality, according to the researchers.

Receiver operating curves analysis yielded an area under the curve value of 0.933, indicating discriminatory ability. With a sensitivity of 92.0% and specificity of 86.0% (P = .001), these metabolic markers could be used to identify patients at heightened risk of disease recurrence and death, potentially informing more personalized and effective treatment strategies, the investigators noted.

Between August 2013 and July 2018, 51 patients with advanced intra-abdominal or pelvic malignancies were referred to the Gynecology Oncology Service at UC Irvine MemorialCare Long Beach Medical Center. Of these patients, 39 had advanced ovarian cancer, 7 had uterine cancer, and 1 had both. Four patients with unrelated malignancies were excluded, leaving 47 patients for evaluation of ex vivo platinum sensitivity and metabolomic analysis. For the control, investigators compared those enrolled with 31 women who were healthy, postmenopausal, aged 54 to 78 years, and with no history of ovarian cancer.

Clinician Perspective

The ability to identify patients with gynecologic cancers, specifically ovarian cancer, who are at highest risk for reoccurrence even after up-front treatment with a platinum-based doublet has eluded researchers for quite some time, explained Sarah Taylor, MD, PhD. Taylor is an assistant professor in the Department of Obstetrics, Gynecology and Reproductive Sciences at the University of Pittsburgh, Pennsylvania.

“We have been able to do prognostication based on stage and optimal cytoreduction, but there is always a subset of individuals who have seemingly good outcomes…and then [develop] platinum-resistant disease. We have yet to develop a good predictor for who those individuals are going to be,” Taylor said.

Taylor further noted that drawing blood prior to treatment to determine the metabolomic signatures associated with this predictor has the potential to be directly clinically applicable, which is an exciting prospect. However, Taylor believes this approach is still very much in its infancy.

“[In the study], there was a limited number of individuals who were included, from an overall cohort perspective, and it was heterogeneous as well,” Taylor noted. “The investigators endeavored to include individuals who had something other than the classic high-grade serous ovary cancer, which I appreciate. However, when you start to include different histologic subtypes for ovarian cancer, endometrial cancer, and even sarcomas, it becomes harder to determine what those responses mean.”

In this study, Taylor explained, researchers collected blood samples and tumor tissue from patients prior to treatment. They analyzed the tumor cells to assess their sensitivity to cisplatin, establishing a median cut-off point to classify patients as either resistant or sensitive to the drug. They then compared these classifications with patient outcomes in real clinical settings. The investigators examined whether the metabolic signatures derived from the blood samples could predict platinum sensitivity or resistance, correlating these findings with the results obtained from the tissue analysis. Taylor advocated for a stronger focus on the practical application of findings, suggesting that relying on blood draws instead of tissue assays would be more beneficial in a clinical context.

For developing a clinically useful predictive signature for platinum sensitivity or resistance, Taylor explained that the focus should be on actual clinical outcomes rather than ex vivo models, which have not proved to be more effective in predicting patient results. Validation of the metabolic signature should rely on its ability to correlate with established clinical outcomes, as these are critical for guiding patient care.

Economical Element

There is an ongoing need for a cost-effective, minimally invasive screening method that can also work well in rural areas, researchers stated in a review published in the Indian Journal of Medical Research.³ The existing methods fall short in these areas, and evaluating the metabolomics with a simple blood sample could lessen the gap in health disparities, the researchers relayed.

Metabolomics has the potential to be a promising economical and accessible option for the early diagnosis of cancer by discovering clinically relevant biomarkers, potentially with just a blood sample. However, challenges remain regarding validation and practical application, and ongoing research may lead to more refined techniques and potentially the ability to screen for multiple cancers through metabolic profiling.

“There is more work that needs to be done to refine its utility, particularly across cancer types, given the small numbers in any one of the given categories [within this study],” Taylor stated. “However, these investigators have what could be an interesting signature.”

REFERENCES:

1. Pang H, Hu Z. Metabolomics in drug research and development: the recent advances in technologies and applications. Acta Pharm Sin B. 2023;13(8):3238-3251. doi:10.1016/j. apsb.2023.05.021

2. D’Amora P, Silva IDCG, Tewari KS, et al. Platinum resistance in gynecologic malignancies: response, disease free and overall survival are predicted by biochemical signature: a metabolomic analysis. Gynecol Oncol. 2021;163(1):162-170. doi:10.1016/j. ygyno.2021.08.001

3. Karekar AK, Dandekar SP. Cancer metabolomics: a tool of clinical utility for early diagnosis of gynaecological cancers. Indian J Med Res. 2021;154(6):787-796. doi:10.4103/ijmr. IJMR_239_19