95 - Genomic Autopsy in Infants with Rare Disease or Sudden Unexpected Death

Poster Number: 95
Publication Number: 95.438

Monica H. Wojcik, Boston Children's Hospital, Wellesley, MA, United States; Sara Vargas, Boston Children's Hospital & Harvard Medical School, Boston, MA, United States; Jill A. Madden, The Manton Center at Boston Children's Hospital, Boston, MA, United States; Casie A. Genetti, Boston Children's Hosptial, Boston, MA, United States; Maya C. del Rosario, Boston Children's Hospital, Boston, MA, United States; Christine Keywan, Boston Children's Hospital, Boston, MA, United States; Bree Martin, Robert's Program - BCH, Boston, MA, United States; Ingrid A. Holm, Harvard Medical School, Boston, MA, United States; Richard D. Goldstein, Boston Children's Hospital, Boston, MA, United States; Pankaj B. Agrawal, Boston Children's Hospital and Harvard Medical School, Boston, MA, United States

Background: Postmortem genomic sequencing, or “genomic autopsy”, is emerging as a powerful tool for understanding conditions leading to death in infancy, though the optimal process for pursuing these diagnoses remains poorly understood.

Objective: To describe our experience incorporating genomics into the postmortem evaluation of infant deaths and the value added by integrating genomic data with the traditional autopsy.

Design/Methods: Prospective cohort study of decedents under 1 year of age with suspected rare disease (RD) or sudden, unexpected death in infancy (SUID). Genomic data and traditional autopsy findings were evaluated and compared.

Results: 359 decedents (316 SUID, 43 RD) were evaluated in a process incorporating referring clinicians or family and our multidisciplinary team. In contrast to SUID, all but 2 RD deaths were anticipated and occurred after planned redirection of care. Sequencing was performed using DNA that was previously saved (4 SUID, 20 RD) or extracted from dried blood spots (11 SUID), peripheral blood (21 SUID, 13 RD), frozen tissue (270 SUID, 3 RD) or formalin-fixed, paraffin-embedded tissue (FFPE, 5 SUID, 2 RD) (5 SUID cases of unknown source). DNA extraction from FFPE tissue was unsuccessful in 6/7 cases, and all other DNA sources proceeded to genomic analysis via exome or genome sequencing. Both conclusive (27/320 SUID, 11/43 RD) and potential (55/320 SUID and 9/43 RD) diagnoses were identified, with analysis ongoing.
Genomic and traditional autopsy both contributed novel information about the underlying disease and cause of death. Three RD cases had diagnoses made via genomic autopsy alone (KMT2D, RET, TBCD). In several cases the traditional autopsy identified causes of death not seen via genomic analysis (e.g., disseminated congenital cytomegalovirus infection) and results suggestive of a non-genetic underlying disorder leading to death (e.g., retroplacental hemorrhage). Genomic autopsy coupled with traditional autopsy offered unique opportunities to explore novel genotype/phenotype correlations down to the microscopic level.Conclusion(s):
We present our approach to genomic autopsy as a powerful tool for postmortem investigation of infant deaths, particularly when accompanied by traditional autopsy. FFPE tissue should be avoided as the sole DNA source. Genomic autopsy and traditional autopsy serve complementary roles to understand infant mortality in situations where the cause of death, underlying diagnosis, or both, are unknown. Together, they allow an unprecedented opportunity to characterize rare disease in addition to providing invaluable information to bereaved parents.

Genomic autopsy processThe process of sample acquisition, consent, and genomic sequencing with results return is illustrated.