Algorithm helps identify which patients can be offered fertility-sparing treatment options

Using the Proactive Molecular Classifier for endometrial cancer (ProMisE) algorithm to classify patients with endometrial cancer or endometrial intraepithelial neoplasia by molecular subtype can help predict which patients will progress following treatment with the levonorgestrel intrauterine (LNG-IUD) system, according to a retrospective analysis.

The levonorgestrel intrauterine system is an alternative to definitive therapy, typically hysterectomy, that is sometimes offered to women with endometrial cancer or neoplasia if they are not good surgical candidates or if they are would like to preserve their fertility. Determining who can safely be given fertility-sparing therapy remains a clinical challenge, however.

This study, which was led by Alison M. Puechl, MD, of Duke University Medical Center in Durham, North Carolina, suggests that patients with endometrial cancer/neoplasia who have an abnormally high copy number of tumor protein 53 (p53) have a shorter time to definitive therapy and/or progression, compared to patients with other molecular subtypes. The findings were to be presented at the Society for Gynecologic Oncology 2020 annual meeting, planned to take place in Toronto, Canada from March 28 to 31. The conference was cancelled due to the Covid-19 pandemic, but the study abstracts were subsequently released.

“As clinicians, we are aware of the limitations of standard pathologic assessment of endometrial cancers,” Tashanna Myers, MD, of Baystate Health in Springfield, Massachusetts, said in a comment on the study for BreakingMED. “Histologic subtype and grade are not sufficient as prognostic markers. Additionally, as we seek to offer patients fertility-sparing options, being able to safely provide those options requires as much information as can be obtained from a sampling of the tumor. Historically, we have used grade and radiologic imaging to help guide those recommendations. As determined by Puechl and her colleagues, there can be significant heterogeneity in the rates of the progression within this cohort of patients with similar histologic diagnoses.”

Using immunohistochemistry and single-gene sequencing, patients were classified into one of four molecular subgroups using the ProMisE algorithm: 1) polymerase-ɛ (POLE) mutated; 2) mismatch repair-deficient (MMR-D); 3) p53 wildtype/POLE wildtype (copy-number low), and 4) p53 abnormal (copy-number high).

“These molecular subtypes have been validated extensively, and the application of ProMisE in this cohort of patients is where we should be going as a subspecialty,” said Meyers.

In total, 48 patients who were treated with the levonorgestrel intrauterine system for medical management of endometrial cancer or endometrial intraepithelial neoplasia from 2013 to 2018 and who had adequate tissue available were included in this retrospective analysis.

Patients’ median age was 55.4 years (range 24-91 years), and the median follow-up time was 16.9 months. Overall, 26 of the 48 patients were diagnosed with endometrial intraepithelial neoplasia prior to levonorgestrel intrauterine system treatment. In total, 20 had grade 1 endometrioid endometrial cancer, and 2 patients had grade 2 disease. Indications for medical management were fertility preservation in 37.5% and medical comorbid conditions precluding surgery in 62.5%.

Overall, 34 patients (71%) were classified as copy-number low, 6 (12.5%) as MMR-D, 4 (8%) as copy-number high, and 4 (8%) as POLE mutated. Among the patients in the copy-number high group, 3 of 4 had endometrial intraepithelial neoplasia with strong and diffuse p53 staining consistent with a missense mutation.

Twelve of the 48 patients included in the analysis (25%) underwent definitive therapy after attempted medical management. Among these patients, 8 (17%) had progressive disease and 4 (8%) had stable disease without regression. The 34 patients who progressed and/or underwent definitive treatment comprised 8 out of 34 (23.5%) of the copy-number low group, 2 out of 6 (33%) of the MMR-D group, 2 out of 4 (50%) of the copy-number high group, and 1out of 4 (25%) of the POLE mutated group.

Patients in the copy-number high group had shortest time to progression and/or definitive therapy, compared with the other three groups (P = 0.03).

“Patient tumors with endometroid endometrial cancer/endometrial intraepithelial neoplasia in the copy-number high subgroup demonstrate a shorter time to definitive therapy/progression when managed with levonorgestrel intrauterine system than tumors in other molecular subgroups,” concluded the authors in their study abstract. “Molecular classification of endometrial cancer/endometrial intraepithelial neoplasia prior to management with levonorgestrel intrauterine system may predict patients at increased risk of progression or needing definitive therapy.”

“This is an important study, as it further supports the biologic differences when categorizing endometrial intraepithelial neoplasia and endometrial cancer by molecular subtype and correlating [that] with response rates to the levonorgestrel intrauterine system,” said Meyers. “As hysterectomy rates decline across the country and patients are seeking fertility-sparing options, understanding those molecular features which place young women at risk for persistent or progressive disease will help to guide treatment decisions.”

She warned, however, that “this was a single institution retrospective study with small numbers, so it cannot be used to guide clinical care. It challenges us to design prospective studies in endometrial cancer that collect and correlate the molecular subtypes. These algorithms should be used in clinical practice; however, we do not yet have guidance on how to modify care once we obtain these data.”

Meyer recommended that investigators “begin to look at these molecular markers and others (e.g., PTEN, PIK3CA, KRAS) in patients choosing fertility-sparing options. However, we should also expand to other subsets of endometrial cancer, such as low-grade tumors with advanced FIGO [Fédération Internationale de Gynécologie et d’Obstétrique] stages and high intermediate risk cancers that may or may not need adjuvant therapy.”

  1. Molecular subtype, using the ProMisE algorithm, can help predict which patients with endometrial cancer or neoplasia are at higher risk of progression following treatment with the levonorgestrel intrauterine system.
  2. Patients with high p53 copy number had a shorter time to progression/definitive therapy when managed with levonorgestrel intrauterine system than patients with other molecular subtypes.

Alison Palkhivala, Contributing Writer, BreakingMED™

Puechl and Meyers reported no conflicts of interest.


Cat ID: 120

Topic ID: 78,120,730,120,445,192,356


Puechl AM, et al “Turning ProMisE into practice: Predicting response in medically managed endometrial cancers via molecular classification” Society of Gynecologic Oncology Annual Meeting on Women’s Cancer 2020; Abstract 29.