Back to 2025 Abstracts

---

Development of the Perihilar Cholangiocarcinoma Risk Estimation of Tumor REcurrence After Transplant (PRETREAT) Score - Insights from 30 Years of Experience
*Zhihao Li, *Timucin Taner, *Sumera Ilyas, *John Eaton, *Ty Diwan, *Gregory Gores, Julie Heimbach
Mayo Clinic, Rochester, MN

Objective
The Mayo protocol (MP) has made liver transplantation (LT) a curative option for unresectable perihilar cholangiocarcinoma (pCCA), but a 30% recurrence rate underscores the need for a risk score to guide surveillance and adjuvant therapy. This study aims to develop the PRETREAT score for patients treated under the MP.
Methods
This retrospective study analyzed all consecutive patients with unresectable pCCA who underwent neoadjuvant chemoradiotherapy followed by LT under the MP at Mayo Clinic Rochester from 1993-2023. Predictors of recurrence were identified using multivariable Cox proportional hazards regression with backward stepwise selection. The PRETREAT score was developed by assigning points based on the model's beta coefficients. Internal validation was performed using 10-fold cross-validation. Survival outcomes were estimated with the Kaplan-Meier method and compared via the log-rank test.
Results
Of 446 patients enrolled in the MP, 297 underwent LT and were included. Among these, 73.7% had pCCA associated with primary sclerosing cholangitis, while the remainder had de novo pCCA. A detectable mass was present in 56.6%, with a median radial diameter of 10mm, and vascular encasement was observed in 24.6%. Deceased and living donor grafts were used in 72.1% and 27.9%. Residual tumor tissue was identified in 50.5% of explants (Table 1).
During a median follow-up of 12.4 years, 89 patients (30.0%) experienced recurrence, with cumulative incidences of 17.1% and 29.2% at 2 and 5 years (Figure 1). Four independent predictors of recurrence were identified: radial diameter per mm (HR 1.03), vascular encasement (HR 2.53), lymphovascular invasion (HR 2.85), and residual tumor (HR 4.83)(Table 2).
The PRETREAT score (range: 0–14) was developed with a c-index of 0.80 (95%CI: 0.71–0.88) and maintained performance in internal validation. Three risk groups were defined: low (0–4), moderate (5–9) and high-risk (10–14). These groups stratified both recurrence-free (RFS) and overall survival (OS)(log-rank p<0.001)(Figure 2). Five-year RFS was 87.6% for low, 50.8% for moderate, and 15.6% for high-risk groups. Corresponding 5-year OS was 88.1%, 61.7%, and 24.2%. Survival outcomes did not differ significantly between 2013–2023 and 1993–2012. The 1/ 3/ 5-year OS was 95.8%/ 82.7%/ 71.4% in 2013–2023, compared to 91.4%/ 77.6%/ 70.3% in 1993–2012. Risk group stratification showed no differences between eras (Figure 3).
Conclusions
Over the past 30 years, the MP has established itself as the standard for LT in patients with unresectable pCCA. While outcomes have been consistently excellent, they have shown no significant improvement over the last decade. The PRETREAT score provides a tool for predicting post-transplant tumor recurrence and survival, effectively stratifying recipients into three risk groups. This stratification offers an opportunity to guide risk-based post-transplant surveillance and tailor adjuvant therapies, potentially enhancing patient outcomes.
Table 1: Demographic and clinicopathological characteristics.

	Overall (n=297)	Recurrence (n=89)	Free of Recurrence (n=208)	P value
Age at transplant [years] (median, IQR)	51 (43-61)	54 (43-62)	50 (42-59)	0.24
Sex male (n, %)	220 (74.1)	57 (64.0)	163 (78.4)	0.02
Primary sclerosing cholangitis (n, %)	219 (73.7)	47 (52.8)	172 (82.7)	<0.001
Mass seen on imaging (n, %)	168 (56.6)	71 (79.8)	97 (46.6)	<0.001
Radial diameter [mm] (median, IQR)	10.0 (0.0-16.3)	14.5 (10.0-19.2)	0.0 (0.0-14.0)	<0.001
Radial diameter ? 30mm (n, %)	6 (2.0)	5 (5.6)	1 (0.5)	0.02
Vascular encasement (n, %)	73 (24.6)	44 (49.4)	29 (13.9)	<0.001
CA 19-9 at transplant [U/mL] (median, IQR)	56 (18-174)	70 (18-219)	48 (18-149)	0.06
<100	188 (63.3)	49 (55.1)	139 (66.8)
100-499	80 (26.9)	27 (30.3)	53 (25.5)
?500	29 (9.8)	13 (14.6)	16 (7.7)
Biological MELD score (median, IQR)	10 (7-17)	10 (7-15)	11 (7-18)	0.13
Graft type (n, %)				0.93
Deceased donor graft	212 (72.1)	65 (73.0)	147 (71.7)
Living donor graft	82 (27.9)	24 (27.0)	58 (28.3)
Concurrent Whipple procedure (n, %)	33 (11.1)	11 (12.4)	22 (10.6)	0.81
Residual tumor tissue on explant (n, %)	150 (50.5)	76 (85.4)	74 (35.6)	<0.001
Lymphovascular invasion (n, %)	23 (7.7)	17 (19.1)	6 (2.9)	<0.001
Perineural invasion (n, %)	58 (19.5)	34 (38.2)	24 (11.5)	<0.001
Time from Mayo protocol enrollment to transplant [months] (median, IQR)	6.2 (3.5-10.7)	6.7 (3.7-10.3)	6.0 (3.5-11.0)	0.62
Transplant era (n, %)				0.99
1993-2012	152 (51.2)	46 (51.7)	106 (51.0)
2013-2023	145 (48.8)	43 (48.3)	102 (49.0)

Abbreviations: CA 19-9, Carbohydrate antigen 19-9; IQR, Interquartile range; MELD, Model for End-Stage Liver Disease. Categorical variables were reported as numbers and percentages, continuous variables as medians and interquartile ranges. Categorical variables were compared using the Chi-square test and continuous variables using the Mann-Whitney U test.
Table 2: Univariate and multivariable Cox proportional hazards regression for posttransplant tumor recurrence and attribution of PRETREAT scores.

	Univariate Cox regression			Multivariable Cox regression			PRETREAT scoring
Predictors	Hazard Ratio	95% CI	P value	Hazard Ratio	95% CI	P value	log(HR)	Points
Age at transplant [years]	1.01	0.99-1.03	0.35
Sex male	0.59	0.38-0.91	0.02
Primary sclerosing cholangitis	0.35	0.23-0.53	<0.001
Mass seen on imaging	3.42	2.04-5.75	<0.001
Radial diameter [mm]	1.07	1.05-1.09	<0.001	1.03	1.00-1.06	0.021	0.0307
<10							0	0
10-20							0.4601	1
20-30							0.7669	2
>30							0.9510	3
Vascular encasement	3.95	2.60-5.99	<0.001	2.53	1.64-3.91	<0.001	0.9283	3
CA 19-9 at transplant (log)	1.12	0.99-1.27	0.064
Living donor graft	0.93	0.58-1.48	0.76
Concurrent Whipple procedure	1.61	0.86-3.03	0.14
Residual tumor tissue on explant	7.79	4.32-14.0	<0.001	4.83	2.60-8.99	<0.001	1.5759	5
Lymphovascular invasion	5.65	3.29-9.70	<0.001	2.85	1.63-4.98	<0.001	1.0466	3
Perineural invasion	3.9	2.53-6.01	<0.001
Time from Mayo protocol enrollment to transplant [months]	0.99	0.95-1.02	0.41

Abbreviations: CA 19-9, Carbohydrate antigen 19-9. Predictors of posttransplant recurrence were identified through backward stepwise elimination, which was guided by the Akaike Information Criterion (AIC). No variables exceeded 5% missing data. Results were expressed as hazard ratios (HR) with 95% confidence intervals (CI). A p<0.05 was considered statistically significant. The PRETREAT risk score-based model (0-14) was developed, with points assigned based on the beta coefficients of the Cox regression model, mirroring the approach used in the Framingham study.
Figure 1: Recurrence dynamics over time. A) Cumulative incidence of recurrence: The 2-year and 5-year cumulative incidences are 17.1% and 29.2%, respectively. B) Temporal distribution of recurrences: The majority of recurrences occur within the first two years post-transplant.