Table 3.

Cancer risks by EIN status and pathology-panel-based WHO classifications

	Cases N	Controls N	RR	95% CI
Panel WHO classification
DPEM	33	97	1.00	Ref.
Non-atypical EH (SH or CH)	62	110	2.24	1.08–4.66
AH	43	34	14.21	5.32–37.96
Collapsed categories:
DPEM or non-atypical EH	95	207	1.00	Ref.
AH	43	34	9.19	3.87–21.83
EIN classification
Benign	71	159	1.00	Ref.
EIN	42	65	7.76	3.36–17.91
Cancer	13	10	17.11	4.18–70.10
Collapsed categories:
Benign	71	159	1.00	Ref.
EIN or cancer	55	75	8.95	4.06–19.74

RRs are based on conditional logistic regression analysis, weighted based on the sampling methods used to select controls. The RRs represent the increased probability of developing carcinoma associated with that EH or EIN classification compared with the probability of develop carcinoma associated with the reference category (i.e., WHO=DPEM, WHO=SH, EIN=benign, or some combination thereof).

Non-atypical EH = SH or CH.

RRs are adjusted for age at index biopsy, date of index biopsy, interval between EH and carcinoma (for cases) or matched censoring date (for controls), BMI at the time of EH diagnosis, MPA-based treatment for EH, and follow-up biopsies.

For EIN classification, 9 cases and 4 controls were judged to be inadequate for EIN classification (RR=2.19, 95% CI, 0.44–10.82). Another 3 case and 3 control slides were not available for review.