Abstract / Summary
Study question. Are endometrial gene expression profiles reproducible across hormone replacement therapy (HRT) cycles and across different regions of the same endometrial biopsy? Summary answer. Endometrial transcriptomic profiles were highly reproducible across HRT cycles and across non-adjacent and adjacent regions of the same biopsy. What is known already. Endometrial transcriptomic analyses are increasingly used to understand endometrial physiology and to evaluate endometrial reproductive potential. However, this approach requires an endometrial biopsy in a prior endometrial cycle to the embryo transfer cycle in the IVF treatment. Therefore, endometrial transcriptomic analyses rely on the assumption that the gene expression profile is representative of subsequent endometrial cycles and of the whole biopsy specimen. However, inter-cycle and intra-biopsy reproducibility remain insufficiently characterized. Study design, size, duration. This was a prospective, single-center, observational clinical study including 20 participants recruited between June 2024 and March 2025. Participants/materials, setting, methods. Participants underwent two different endometrial preparation cycles under the same HRT conditions at a private fertility clinic, and an endometrial biopsy was collected during both cycles, with the same duration of progesterone exposure, in the mid-secretory phase by the same clinician. RNA was processed for whole endometrial transcriptome analysis using RNA sequencing. Three paired comparisons were performed: biopsies from the same participant collected in different cycles, non-adjacent regions of the same biopsy, and adjacent regions of the same biopsy. Gene expression reproducibility was assessed using Pearson correlation, hierarchical clustering, principal component analysis, and paired differential gene expression analysis. The reproducibility of estimated cell-type proportions was also assessed by deconvolution to observe potential differences in cellular composition. Main results and the role of chance. In the inter-cycle endometrial gene expression comparison, the mean {+/-} standard deviation (SD) gene expression correlation coefficient within participant pairs was 0.974 {+/-} 0.012, significantly higher than that observed between participant pairs (p = 1.56 x 10-6), the mean {+/-} SD Euclidean distance was significantly lower within than between participant pairs (p = 2.24 x 10-6), and no differentially expressed genes were identified within participant pairs. Principal component analysis and hierarchical clustering further supported the similarity of most paired samples. Intra-biopsy comparisons showed even higher reproducibility, with both non-adjacent and adjacent biopsy regions showing mean {+/-} SD correlation coefficients of 0.986 {+/-} 0.003 and 0.986 {+/-} 0.002, respectively. The correlation coefficients were significantly higher than those observed between biopsy pairs (p = 9.91 x 10-12 and p = 2.14 x 10-12, respectively). No differentially expressed genes were detected within intra-biopsy pairs, all paired regions clustered together by hierarchical clustering and principal component analysis, and intra-biopsy pairs showed lower Euclidean distances than inter-biopsy pairs (p = 9.91 x 10-12 and p = 2.14 x 10-12 in non-adjacent and adjacent regions, respectively). Estimated cell-type proportions did not differ significantly between paired samples in any of the three comparisons. Limitations, reasons for caution. The findings are specific to samples analyzed using the AmpliSeq Transcriptome Human Gene Expression Panel protocol. Although other transcriptomic platforms should be tested, we are expecting the same reproducibility due to the consistency of RNA seq platforms. All biopsies were collected by the same clinician, so operator-related variability could not be directly assessed. These findings can be extrapolated only to HRT cycles. Wider implications of the findings. These findings support the reproducibility of endometrial gene expression profiles under controlled HRT conditions from biopsies collected by the same clinician and suggest that a single biopsy region can provide a representative molecular profile of the whole biopsy specimen. These findings strengthen the methodological basis for bulk endometrial RNA sequencing studies and support the reliability of transcriptomic-based tools for endometrial evaluation in an endometrial cycle prior to an embryo transfer cycle for reproductive medicine.