Table 3.

Assay’s performance in detecting MRD in the evaluable population, excluding the patient with metastatic disease during treatment.

	Assay	N	N Relapse	N Positive	N True positive	NPV (%)	PPV (%)	Sensitivity (%)	Specificity (%)	Accuracy (%)
FU1	RaDaR	14	5	0	0	64.3	–	0	100	64.3
	CAPP-seq	24	5	4	2	85	50	40	89.5	79.2
	HPV-seq	15	4	3	2	83.3	66.7	50	90.9	80
	dPCR	15	4	2	2	84.6	100	50	100	86.7
FU2	RaDaR	16	5	2	2	78.6	100	40	100	81.3
	CAPP-seq	26	5	3	1	82.6	33.3	20	90.5	76.9
	HPV-seq	14	3	3	3	100	100	100	100	100
	dPCR	14	3	1	1	84.6	100	33.3	100	85.7
Any FU	RaDaR	17	6	2	2	73.3	100	33.3	100	76.4
	CAPP-seq	29	6	4	2	84	50	33.3	91.3	79.3
	HPV-seq	16	4	5	4	100	80	100	91.7	93.8
	dPCR	16	4	2	2	85.7	100	50	100	87.5
Combination of different assays
FU1	All Assays	13	4	5	3	77.8	75	60	87.5	76.9
	RaDaR + HPV-seq	13	5	3	3	80	100	60	100	84.6
	CAPP-seq + HPV-seq	24	5	5	2	84.2	40	40	84.2	75
FU2	All Assays	16	5	5	4	90.9	80	80	91	87.5
	RaDaR + HPV-seq	16	5	4	4	91.7	100	80	100	93.8
	CAPP-seq + HPV-seq	25	5	5	3	90	60	60	90	84
Any FU	All Assays	14	6	5	5	87.5	83.3	83.3	87.5	85.7
	RaDaR + HPV-seq	14	6	5	5	88.9	100	83.3	100	92.3
	CAPP-seq + HPV-seq	21	5	7	4	92.3	57.1	80	81.3	81

Patients with a sample not available for analysis were not considered for analysis at the specific follow-up, while a patient with one follow-up sample negative but the other not performed/available, were removed from the analysis at any follow-up.

FU follow up, NPV negative predictive value, PPV positive predictive value.