Long term longitudinal follow-up of an AD-HIES cohort: the impact of early diagnosis and enrollment to IPINet centers on the natural history of Job's syndrome

Obstructive uropathy in STAT 3 hyper immunoglobulin E syndrome: A 5 year old Middle Eastern boy

Autosomal dominant hyper immunoglobulin E (IgE) syndrome is a rare inborn error of immunity that affects approximately one in a million individuals worldwide. It presents with various symptoms owing to multisystem involvement (immunological and non-immunological). Recurrent infections (mainly in the skin and lungs) are common presentations. A 5-year-old Middle Eastern boy presented with symptoms suggestive of obstructive uropathy secondary to multiple large pelviabdominal abscesses and acute kidney injury with hyperkalemia that necessitated admission to the intensive care unit. Upon further investigation, the patient's genetic test (whole exome sequencing) demonstrated a heterozygous missense variant in the STAT3 gene. The patient completely recovered and did not require further admission after initiating prophylactic antibiotics. Although deep-seated infections are uncommon in STAT3 hyper IgE syndrome, skin and lung infections are most commonly observed. Multiple deep collections can occur and require prompt intervention and aggressive treatment.

Natural language processing of clinical notes enables early inborn error of immunity risk ascertainment

Background

There are now approximately 450 discrete inborn errors of immunity (IEI) described; however, diagnostic rates remain suboptimal. Use of structured health record data has proven useful for patient detection but may be augmented by natural language processing (NLP). Here we present a machine learning model that can distinguish patients from controls significantly in advance of ultimate diagnosis date.

Objective

We sought to create an NLP machine learning algorithm that could identify IEI patients early during the disease course and shorten the diagnostic odyssey.

Methods

Our approach involved extracting a large corpus of IEI patient clinical-note text from a major referral center's electronic health record (EHR) system and a matched control corpus for comparison. We built text classifiers with simple machine learning methods and trained them on progressively longer time epochs before date of diagnosis.

Results

The top performing NLP algorithm effectively distinguished cases from controls robustly 36 months before ultimate clinical diagnosis (area under precision recall curve > 0.95). Corpus analysis demonstrated that statistically enriched, IEI-relevant terms were evident 24+ months before diagnosis, validating that clinical notes can provide a signal for early prediction of IEI.

Conclusion

Mining EHR notes with NLP holds promise for improving early IEI patient detection.