Giulia KennedyPhD
PinkDxChapter 14 · Strand B · Research

Reading endometrial cancer from a vaginal swab

The PNK001 study showed that molecular signals of endometrial cancer can be read from a vaginal swab by whole-transcriptome sequencing and machine learning, published in April 2026; the PROACTION study is validating the test.

JUL 2024 – PRESENTCO-FOUNDER AND CHIEF SCIENTIFIC OFFICER, PINKDXPINKDX, INC. · DALY CITY, CALIFORNIA

Overview

“The assumption has always been that you have to sample the uterus to understand what’s happening in the uterus,” Kennedy said when PinkDx published its first peer-reviewed study on April 6, 2026. “Our findings challenge that assumption. We show that molecular signals associated with endometrial cancer can be detected from a vaginal swab.”1

The study

The evidence came from PNK001, an exploratory study registered as NCT06460818 that enrolled 236 participants between February and November 2024 at Mayo Clinic, Columbia University and Houston Methodist.23 Women undergoing hysterectomy provided vaginal swabs, ectocervical swabs, endocervical cytobrushes and endometrial tissue, so that each less invasive sample could be compared with the tissue diagnosis.4

Whole-transcriptome sequencing was run on samples from 27 PNK001 participants and 46 Cooperative Human Tissue Network specimens. Classifiers trained on expressed genes and variant counts distinguished 5 benign from 15 malignant cases in cytobrush, ectocervical and vaginal swab samples, with average cross-validation AUCs of 0.6 to 0.96; tissue-trained classifiers reached AUCs of 0.97 and 0.98 on an independent test set.42 The paper appeared in the International Journal of Gynecological Cancer with Kennedy among the authors.

Toward a test

Validation is running in PROACTION (PNK002, NCT06527157), an observational study that began on July 30, 2024 and is being conducted at 16 U.S. sites chosen to build diversity into the cohort, in women aged 45 and older with abnormal uterine or postmenopausal bleeding.56 The company expects to validate its first product by the end of 2026 and to launch it by mid-2027 as a laboratory-developed test.67 The work was featured on ABC’s Good Morning America in April 2026, which reported clinical validation under way at 12 major medical centres including Mayo Clinic and Columbia University.7

The thread back to Afirma is direct: a classifier trained on high-dimensional RNA data, validated prospectively, in place of a procedure.

From the record

Facts

  • The exploratory PNK001 study (NCT06460818) enrolled 236 participants between February and November 2024 at Mayo Clinic, Columbia University and Houston Methodist; women undergoing hysterectomy provided vaginal swabs, ectocervical swabs, endocervical cytobrushes and endometrial tissue.243
  • On April 6, 2026 PinkDx announced publication in the International Journal of Gynecological Cancer of evidence that molecular signals associated with endometrial cancer can be detected from a vaginal swab using whole-transcriptome sequencing and machine learning; Kennedy is a co-author and was quoted as co-founder and chief scientific officer.12
  • In the published study, classifiers trained on expressed genes and variant counts distinguished 5 benign from 15 malignant cases in cytobrush, ectocervical and vaginal swab samples with average cross-validation AUCs of 0.6 to 0.96; sequencing data came from 27 PNK001 participants and 46 Cooperative Human Tissue Network samples.42
  • PinkDx sponsors the PROACTION study (PNK002, NCT06527157), an observational study that began July 30, 2024 across 16 U.S. sites in women aged 45 and older with abnormal uterine or postmenopausal bleeding.56
  • PinkDx expects to validate its first product by the end of 2026 and launch it by mid-2027 as a laboratory-developed test, with clinical validation under way at 12 major medical centres including Mayo Clinic and Columbia University.67
  • The research was featured on ABC's Good Morning America in April 2026, with Kennedy quoted as co-founder and chief scientific officer.7

Contributions

  • Co-author of the first peer-reviewed evidence that endometrial cancer signals can be detected from a non-invasive vaginal swab (International Journal of Gynecological Cancer, 2026).14

Research

  • Gynecologic cancer detection

    2022 – PRESENT

    At PinkDx, co-founded with Bonnie Anderson and launched publicly on April 25, 2024 with a $40 million Series A, Kennedy leads scientific strategy and translational research applying whole-transcriptome sequencing and machine learning to gynecologic cancer, starting with endometrial cancer. The exploratory PNK001 study enrolled 236 women undergoing hysterectomy in 2024 and showed that molecular signals of endometrial cancer can be detected from a vaginal swab (International Journal of Gynecological Cancer, April 2026); the PROACTION study (NCT06527157) is validating the test across 16 U.S. sites.

  • Machine learning for clinical diagnostics

    DEC 2010 – PRESENT

    The method that links every classifier Kennedy has helped build since 2010: training algorithms on high-dimensional genomic data from clinical samples and validating them against histopathology. She presented on Veracyte's use of machine learning to translate genomic data into commercial diagnostic tests at the Precision Medicine World Conference in January 2018, and described the Afirma GSC's ensemble machine learning algorithms as deriving "from years of thoughtful scientific research, data analysis and statistical modeling". The same approach underlies the Envisia and Percepta classifiers and PinkDx's whole-transcriptome vaginal swab test.

Technologies
  • PinkDx vaginal swab classifier for endometrial cancer
  • Whole-transcriptome RNA sequencing
  • Ensemble machine learning classifiers

Publications