Hypofractionated radiotherapy-related lymphopenia is associated with worse survival in unresectable intrahepatic cholangiocarcinoma

Abstract

Objective:

The aim of this study was to evaluate the incidence of radiotherapy (RT)-related lymphopenia, its predictors, and association with survival in unresectable intrahepatic cholangiocarcinoma (ICC) treated with hypofractionated-RT (HF-RT).

Methods:

Retrospective analysis of 96 patients with unresectable ICC who underwent HF-RT (median 58.05 Gy in 15 fractions) between 2009–2022 was performed. Absolute lymphocyte count (ALC) nadir within 12 weeks of RT was analyzed. Primary variable of interest was severe lymphopenia, defined as Grade 3+ (ALC <0.5 k/μl) per CTCAE v5.0. Primary outcome of interest was overall survival (OS) from RT.

Results:

Median follow-up was 16 months. 52% of patients had chemotherapy pre-RT, 23% during RT, and 40% post-RT. Pre-RT, median ALC was 1.1 k/μl and 5% had severe lymphopenia. Post-RT, 68% developed RT-related severe lymphopenia. Patients who developed severe lymphopenia had a significantly lower pre-RT ALC (median 1.1 vs 1.5 k/μl, p=0.01) and larger target tumor volume (median 125 vs 62 cc, p=0.02). In our multivariable Cox model, severe lymphopenia was associated with a 1.7-fold increased risk of death (p=0.04); 1-year OS rates were 63% vs 77% (p=0.03). Receipt of photon vs proton-based RT (OR=3.50, p=0.02), higher mean liver dose (OR=1.19, p<0.01), and longer RT duration (OR=1.49, p=0.02) predicted severe lymphopenia.

Conclusion:

HF-RT related lymphopenia is an independent prognostic factor for survival in patients with unresectable ICC. Patients with lower baseline ALC and larger tumor volume may be at increased risk, and use of proton therapy, minimizing mean liver dose, and avoiding treatment breaks may reduce RT-related lymphopenia.

Introduction

Intrahepatic cholangiocarcinoma (ICC) is a lethal malignancy that arises from the intrahepatic bile duct and accounts for 12% of primary hepatic tumors.¹ For localized, resectable tumors, surgery is a potentially curative treatment. However, approximately 70% of cases are unresectable^2,3 and are associated with a poor median survival of only 7 to 12 months.⁴ For unresectable ICC, palliative chemotherapy with gemcitabine/cisplatin⁵ and more recently gemcitabine/cisplatin/durvalumab⁶ is the standard of care. Studies also suggest that delivery of high, ablative doses of radiotherapy (RT) is safe and can improve local control and survival in unresectable ICC.^4,7–10 High-dose hypofractionated RT (HF-RT) has been established as a potential definitive treatment for localized, unresectable ICC by a phase II, multi-institutional study on the use of proton beam therapy for primary hepatic tumors.¹¹

While RT can improve tumor control and survival, it often leads to depletion of the circulating lymphocytes^12–14 which are key mediators of the host anti-tumor immune response.¹⁵ In multiple solid malignancies including ICC, absolute lymphocyte counts (ALC) and/or tumor-infiltrating lymphocytes (TIL) have been identified as important prognostic factors.^16–18 Prior retrospective studies in glioblastoma, head and neck cancer, non-small cell lung cancer (NSCLC), esophageal cancer, pancreatic adenocarcinoma, cervical, and anal cancer have demonstrated associations between RT-related lymphopenia and worse survival.^19–25 Studies of liver-directed RT in hepatocellular carcinoma (HCC) also suggest an association between severe lymphopenia and worse survival.^26–30 Given the liver’s rich blood circulation, a large pool of circulating lymphocytes is likely exposed to radiation during treatment of hepatic tumors, with a potential for development of severe lymphopenia. To date, RT-related lymphopenia and its impact on the survival outcome of unresectable ICC have not been explored. With the growing incidence of ICC^31,32 and the increasing use of ablative RT as well as immunotherapy for the management of unresectable ICC, identification of various risk factors for RT-related lymphopenia and its prognostic impact will be crucial.

We hypothesized that a subset of patients undergoing HF-RT for unresectable ICC would develop lymphopenia and severe lymphopenia would be prognostic of worse survival. Accordingly, in this study, we evaluated RT-related lymphopenia, its predictors, and association with survival in patients who underwent HF-RT for unresectable ICC.

Materials and Methods

Patients

This retrospective study was conducted under institutional review board approval. Between 2009 and 2022, we identified 96 patients with unresectable ICC who underwent HF-RT at our institution. All patients who received proton therapy were also enrolled in our previously published phase II clinical trial (NCT00976898) of protons in unresectable primary liver tumors.¹¹ Patient selection criteria were as previously explained.^10,11 Briefly, the criteria for patients enrolled in the clinical trial were: 1) one to three primary liver tumors (<12 cm for solitary tumors), 2) absence of extrahepatic disease, and 3) Child-Turcotte-Pugh (CTP) class of A or B. Patients who had lymphopenia prior to HF-RT were not excluded. For non-clinical trial patients, there were no strict criteria but patients who had cirrhosis with CTP class C were excluded. Patients who were not evaluable for the primary variable of interest (ALC nadir within 12 weeks of RT) and/or lost to follow-up were excluded.

Patient and disease characteristics at time of RT including age, sex, race, Eastern Cooperative Oncology Group (ECOG) performance status, immunosuppression status (i.e., human immunodeficiency virus [HIV], transplant, or splenectomy), number of liver lesions present and treated, maximum diameter of treated tumor, presence of cirrhosis, biliary obstruction, vascular invasion, extrahepatic metastasis, tumor markers, and liver function test values were collected. Baseline pre-RT blood counts including white blood cell (WBC), hemoglobin (HGB), platelet (PLT), absolute neutrophil count (ANC), and ALC were recorded. Systemic therapy details including chemotherapy prior to, during, and following RT as well as interval with respect to RT were abstracted.

Radiotherapy

RT planning and delivery were as previously described.^10,11 Briefly, all patients underwent 4-dimensional (4D) simulation with oral and intravenous contrast. Gross total volume (GTV) was contoured based on imaging with magnetic resonance imaging (MRI) liver fusion when available. Clinical target volume (CTV) was generated with a 3–10 mm expansion at the discretion of the treating physician. Planning target volume (PTV) was generated with a 5–10 mm expansion according to the institutional motion management protocol. Daily pretreatment verification was performed with kV imaging for proton therapy and cone beam computed tomography (CBCT) for photon therapy. Prescription dose determined by V_effective liver dose constraints was delivered in 15 daily fractions. Photon treatments were delivered with intensity-modulated RT (IMRT) or volumetric arc therapy (VMAT). Proton treatments were delivered with 240 MeV cyclotron.

Lymphocyte counts

Baseline ALC prior to RT and ALC nadir within 12 weeks of initiating RT were recorded. The primary variable of interest was ALC nadir since initiating RT, stratified by presence vs absence of severe lymphopenia, which was defined as Grade 3 or worse lymphopenia (ALC <0.5 k/μl). Lymphopenia was classified according to the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0.³³

Statistical analysis

Descriptive statistics were used to summarize baseline clinical and treatment characteristics. Kruskal-Wallis and Fisher’s exact tests were used to compare continuous and categorical variables, respectively, between patients who did and did not develop severe lymphopenia. Univariable and multivariable logistic regression analyses were performed to evaluate variables associated with risk of developing severe lymphopenia. The primary endpoint of the study was overall survival (OS), calculated from the time of RT to death of any cause or last follow up. The Kaplan-Meier method was used to calculate the actuarial rates of death. The log-rank test was used to compare the survival probabilities. Associations between clinical and treatment variables with OS were assessed by univariable and multivariable Cox proportional hazards regression models. Statistical significance was defined as a two-sided P value <0.05. SAS 9.4 was used to perform all statistical analyses.

Results

Patient characteristics

A total of 96 patients with unresectable ICC who underwent HF-RT at our institution were evaluated. Baseline patient characteristics are summarized in Table 1. Median age was 73 (range 29–91). ECOG performance status was 0 in 53% (n=51). At the time RT, 9% (n=9) were immunosuppressed: 6% (n=6) had HIV, 1% (n=1) had a history of transplantation, and 2% (n=2) had history of splenectomy. Twenty-four percent (n=23) had extrahepatic disease. Total number of intrahepatic tumors present were one in 76% (n=73), two in 10% (n=10), and three or greater in 14% (n=13). Median tumor size was 5.6 cm (range 2.3–16.0) in the largest dimension. At the time of RT, 9% (n=9) had cirrhosis, 21% (n=20) had biliary obstruction, and 27% (n=26) had macrovascular invasion. The median CEA and CA19–9 at the time of RT were 2.5 ng/ml (range 0.5–40.7) and 72 U/ml (range <0.2–10,549), respectively.

Table 1.

Patient characteristics

Characteristics	All patients (N=96)	No severe lymphopenia* (N=31)	Severe lymphopenia* (N=65)	P value
Patient characteristics
Age, median (range)	73 (29–91)	76 (29–85)	70 (36–91)	0.143
Sex, N(%)				0.276
Female	47 (49)	18 (42)	29 (45)
Male	49 (51)	13 (58)	36 (55)
Race/ethnicity, N(%)				>0.999
Caucasian	85 (89)	28 (90)	57 (88)
African American	2 (2)	1 (3)	1 (2)
Asian	5 (5)	1 (3)	4 (6)
Other	4 (4)	1 (3)	3 (5)
ECOG PS, N(%)				0.382
0	51 (53)	14 (45)	37 (57)
1–4	45 (47)	17 (55)	28 (43)
Immunosuppression, N(%)				0.078
None	87 (91)	29 (94)	58 (89)
HIV	6 (6)	0 (0)	6 (9)
Other (transplant, splenectomy)	3 (3)	2 (6)	1 (2)
Baseline CBC, median(IQR)
WBC (k/μL)	6.9 (4.8–8.5)	7.3 (5.0–9.6)	6.8 (4.7–8.0)	0.413
HGB (g/dL)	11.4 (10.2–12.7)	11.6 (10.8–12.6)	10.9 (10.0–12.8)	0.378
PLT (k/μL)	198 (123–255)	214 (156–284)	186 (117–248)	0.102
ANC (k/μL)	4.5 (3.0–5.9)	5.0 (3.5–6.0)	4.4 (2.0–5.8)	0.548
ALC (k/μL)	1.1 (0.8–1.6)	1.5 (1.0–2.1)	1.1 (0.8–1.3)	0.005
Disease characteristics
Number of tumors, N(%)				0.401
1	73 (76)	27 (87)	46 (71)
2	10 (10)	2 (6)	8 (12)
3	4 (4)	1 (3)	3 (5)
≥4	9 (9)	1 (3)	8 (12)
Number of tumors treated, N(%)				0.338
1	86 (90)	30 (97)	56 (86)
2	8 (8)	1 (3)	7 (11)
3	2 (2)	0 (0)	2 (3)
Tumor size (cm), median(IQR)	5.6 (4.0–8.0)	5.1 (4.2–7.9)	6.0 (4.0–8.0)	0.385
Extrahepatic metastasis, N(%)	23 (24)	5 (16)	18 (28)	0.307
Cirrhosis, N(%)	9 (9)	2 (7)	7 (11)	0.714
Liver function test, median(IQR)
Albumin (g/dl)	4.0 (3.5–4.3)	4.0 (3.5–4.3)	3.9 (3.5–4.3)	0.762
Total bilirubin (mg/dl)	0.5 (0.3–1.1)	0.5 (0.3–0.9)	0.6 (0.4–1.1)	0.243
INR	1.1 (1.0–1.2)	1.1 (1.0–1.2)	1.1 (1.0–1.2)	0.476
Tumor marker, median(IQR)
CEA (ng/ml)	2.5 (1.6–4.5)	2.5 (1.6–4.0)	3.1 (2.4–5.6)	0.367
CA19–9 (U/ml)	72 (23–210)	38 (156–127)	76 (24–221)	0.453
Biliary obstruction, N(%)	20 (21)	5 (16)	15 (23)	0.593
Macrovascular invasion, N(%)	26 (27)	8 (26)	18 (28)	>0.999
Treatment characteristics
Chemo prior to RT, N(%)	50 (52)	18 (58)	32 (49)	0.514
Gemcitabine/cisplatin	29 (58)	9 (50)	20 (63)
Gemcitabine/oxaliplatin	11 (22)	5 (28)	6 (19)
FOLFOX	6 (12)	2 (11)	4 (13)
Others	4 (8)	2 (11)	2 (6)
Last chemo prior to RT (days), median(range)	41 (11–221)	44 (21–134)	41 (11–221)	0.384
Chemo during RT, N(%)	22 (23)	6 (19)	16 (25)	0.615
5-fluorouracil	5 (23)	1 (17)	4 (25)	0.744
Capecitabine	17 (77)	5 (83)	12 (75)
Chemo post RT, N(%)	38 (40)	14 (45)	24 (37)	0.506
Gemcitabine/cisplatin	10 (26)	4 (29)	6 (25)
Gemcitabine/oxaliplatin	3 (8)	2 (14)	1 (4)
FOLFOX	8 (21)	0 (0)	8 (33)
Gemcitabine alone	5 (13)	3 (21)	2 (8)
Gemcitabine/cisplatin/durva	4 (11)	3 (21)	1 (4)
Others	8 (21)	2 (14)	6 (25)
First chemo post RT (days), median(range)	186 (15–1169)	184 (48–1169)	186 (15–702)	0.372

^*Severe lymphopenia defined as absolute lymphocyte count (ALC) <0.5k/μl within 12 weeks of initiating radiotherapy (RT)

Abbreviations: RT=radiotherapy, aLC=absolute lymphocyte count, ECOG PS=Eastern Cooperative Oncology Group performance status, HIV=human immunodeficiency virus, CBC=complete blood count, IQR=interquartile range, WBC=white blood cell, HGB=hemoglobin, PLT=platelet, ANC=absolute neutrophil count, FOLFOX=5-fluorouracil/oxaliplatin/leucovorin

In this study cohort, 68% (n=65) developed severe lymphopenia (ALC <0.5 k/μl) post-RT while 32% (n=31) did not. There were no significant differences in baseline patient and disease characteristics including age, sex, race, performance status, immunosuppression, baseline WBC, HGB, PLT, ANC, number of tumors, tumor size, liver function tests, tumor markers, presence of cirrhosis, biliary obstruction, macrovascular invasion, or extrahepatic metastasis between the cohort who developed vs did not develop severe lymphopenia.

Treatment characteristics

Chemotherapy details are summarized in Table 1. Prior to RT, 52% (n=50) underwent chemotherapy with last chemotherapy date at median of 41 days (range 11–221; 22% [n=11] within 30 days) from start of RT; common chemotherapy regimens included gemcitabine/cisplatin (58%, n=29), gemcitabine/oxaliplatin (22%, n=11), and FOLFOX (5-fluorouracil/oxaliplatin/leucovorin; 12%, n=6); only one patient had received gemcitabine/cisplatin/durvalumab. Twenty-three percent (n=22) received concurrent chemotherapy with capecitabine (77%, n=17) or 5-fluorouracil (23%, n=5). Forty percent (n=38) received chemotherapy following RT at median of 186 days (range 15–1169; 18% [n=7] within 60 days) since completion of RT; common chemotherapy regimens included gemcitabine/cisplatin (26%, n=10), FOLFOX (21%, n=8), gemcitabine alone (13%, n=5), gemcitabine/cisplatin/durvalumab (11%, n=4), and gemcitabine/oxaliplatin (8%, n=3). There was no significant difference in receipt of chemotherapy prior to, during, or after RT between patients who developed vs did not develop severe lymphopenia. Out of five patients (5%) who had severe lymphopenia at baseline prior to RT, only one patient had received chemotherapy prior to (gemcitabine/cisplatin).

Table 2 summarizes the RT characteristics. The median time from initial diagnosis to RT was 123 days (range 7–2260). RT was photon-based in 70% (n=67) and proton-based in 30% (n=29). The median RT dose was 58.05 Gy (range 37.5–67.5), all delivered in 15 daily fractions. The median BED was 80.5 Gy (range 46.9–97.9). The median cumulative GTV was 96 cc (range 4–1172). The median liver volume was 1731 cc (range 687–3280) and the median mean liver dose (MLD) was 22.3 Gy (range 6.4–31.7). Patients who developed severe lymphopenia had a shorter time from initial diagnosis to RT (median 110 vs 192 days, p=0.042), were less likely to have received proton therapy (22% vs 48%, p=0.010), and had a longer duration of RT (median 21 vs 20 days, p=0.011; >21 days in 28% vs 3%, p=0.005), greater cumulative GTV (median 125 vs 62 cc, p=0.015), and higher MLD (23.5 vs 19.5 Gy, p<0.001).

Table 2.

Radiotherapy characteristics

Parameters	All patients (N=96)	No severe lymphopenia* (N=31)	Severe lymphopenia* (N=65)	P value
Diagnosis to RT (days), median(range)	123 (7–2260)	192 (31–1148)	110 (7–2260)	0.042
Radiation modality, N(%)				0.010
Photon	67 (70)	16 (52)	51 (78)
Proton	29 (30)	15 (48)	14 (22)
RT duration (days), median(range)	21 (18–38)	20 (18–22)	21 (18–38)	0.011
RT duration >21 days, N(%)	19 (20)	1 (3)	18 (28)	0.005
RT dose (Gy), median(IQR)	58.1 (52.5–67.5)	58.1 (52.5–67.5)	58.1 (52.5–60.4)	0.496
BED (Gy), median(IQR)	80.5 (70.9–97.9)	80.5 (70.9–97.9)	80.5 (70.9–97.9)	0.496
Fraction, median(IQR)	15 (15–15)	15 (15–15)	15 (15–15)	--
Total GTV (cc), median(IQR)	96 (50–216)	62 (44–134)	125 (61–233)	0.015
Liver volume (cc), median(IQR)	1731 (1331–2059)	1746 (1271–2110)	1727 (1384–2045)	0.903
Mean liver dose (Gy), median(IQR)	22.3 (18.5–26.1)	19.5 (14.0–22.5)	23.5 (20.2–26.7)	<0.001

^*Severe lymphopenia defined as absolute lymphocyte count (ALC) <0.5k/μl within 12 weeks of initiating radiotherapy (RT)

Abbreviations: RT=radiotherapy, ALC= absolute lymphocyte count, IQR=interquartile range, BED=biologic effective dose, GTV=gross target volume

Lymphocyte counts

The median baseline ALC was 1.1 k/μl (range 0.2–6.4) for the entire cohort. Patients who developed severe lymphopenia had slightly lower baseline ALC prior to RT (median 1.1 vs 1.5 k/μl, p=0.01). Five (5%) patients had Grade 3 lymphopenia prior to RT; among these patients, two developed Grade 4 lymphopenia and three continued to have Grade 3 lymphopenia. The median ALC nadir within 12 weeks of initiating RT was 0.4 k/μl (range 0–1.5) at median time of 20 days (IQR 15–29) since start of RT for the entire cohort. The majority (75%) of patients experienced an ALC nadir within 30 days of starting RT. Sixty-eight percent (n=65) developed severe lymphopenia (ALC <0.5 k/μl) and among them 13 patients (14% of entire cohort) had Grade 4 (ALC <0.2 k/μl) lymphopenia. The median ALC nadir was 0.3 k/μl (range 0–0.49) for the cohort who developed severe lymphopenia vs 0.7 k/μl (range 0.5–1.5) for the cohort who did not develop severe lymphopenia. Among patients who did not develop severe lymphopenia, 4 (13%) had no lymphopenia while 5 (16%) had Grade 1 and 22 (71%) had Grade 2 lymphopenia.

Predictors of severe lymphopenia

On univariable logistic regression analysis, receipt of photon vs proton-based RT (OR=3.42 [95% CI 1.36–8.57], p=0.009), higher MLD (OR=1.17 [95% CI 1.07–1.28], p<0.001), and longer RT duration (OR=1.41 [95% CI 1.06–1.87], p=0.018) were associated with increased likelihood of developing severe lymphopenia (Table 3). On multivariable logistic regression model incorporating all significant variables, receipt of photon vs proton-based RT (OR=3.50 [95% CI 1.21–10.11], p=0.020), higher MLD (OR=1.19 [95% CI 1.08–1.32], p<0.001), and longer RT duration (OR=1.49 [95% CI 1.07–2.09], p=0.020) all remained statistically significantly associated with increased likelihood of developing severe lymphopenia. Other variables such as patient and tumor characteristics including immunosuppression and cirrhosis, baseline labs including ALC, receipt of chemotherapy including within 30–60 days of RT and chemotherapy regimen, prescription RT dose, and tumor target volume were not associated with severe lymphopenia.

Table 3.

Univariable and multivariable logistic regression analysis for predictors of severe lymphopenia^*

	Univariable		Multivariable
Variables	OR (95% CI)	P value	OR (95% CI)	P value
Age (years)	0.98 (0.94–1.02)	0.244	--	--
Sex, male vs female	1.72 (0.72–4.08)	0.220	--	--
ECOG PS ≥1 vs 0	0.62 (0.26–1.47)	0.282	--	--
Immunosuppression	1.75 (0.34–8.96)	0.502	--	--
Baseline WBC (k/μl)	1.03 (0.96–1.11)	0.461	--	--
Baseline HGB (g/dL)	0.91 (0.72–1.16)	0.457	--	--
Baseline PLT (k/μl)	1.00 (0.99–1.00)	0.091	--	--
Baseline ANC (k/μl)	1.03 (0.92–1.16)	0.609	--	--
Baseline ALC (k/μl)	0.62 (0.34–1.11)	0.109	--	--
Baseline ALC <1	1.67 (0.66–4.23)	0.283	--	--
Number of tumors treated
1	Ref	--	--	--
2	3.75 (0.44–31.93)	0.226	--	--
Tumor size (cm)	1.07 (0.91–1.25)	0.430	--	--
Extrahepatic metastasis	1.99 (0.66–5.99)	0.220	--	--
Cirrhosis	1.75 (0.34–8.96)	0.502	--	--
Baseline albumin (g/dL)	0.84 (0.39–1.83)	0.661
Baseline INR	1.04 (0.36–3.00)	0.943	--	--
Baseline CEA (ng/mL)	1.01 (0.93–1.10)	0.806
Baseline CA19–9 (U/mL)	1.00 (1.00–1.00)	0.611
Biliary obstruction	1.56 (0.51–4.77)	0.436
Macrovascular invasion	1.10 (0.42–2.91)	0.846	--	--
Any chemo	0.56 (0.22–1.44)	0.225
Chemo prior to RT	0.70 (0.30–1.66)	0.419	--	--
Gemcitabine/cisplatin	Ref	--	--	--
Gemcitabine/oxaliplatin	0.54 (0.13–2.24)	0.396	--	--
FOLFOX	0.90 (0.14–5.84)	0.912	--	--
Others	0.45 (0.05–3.72)	0.459	--	--
Chemo within 30 days before RT	1.31 (0.32–5.32)	0.706	--	--
Gemcitabine/cisplatin	Ref	--	--	--
Others	0.67 (0.04–11.94)	0.783	--	--
Chemo within 60 days before RT	0.87 (0.36–2.07)	0.745	--	--
Gemcitabine/cisplatin	Ref	--	--	--
Gemcitabine/oxaliplatin	0.41 (0.05–3.53)	0.418	--	--
FOLFOX	0.82 (0.12–5.57)	0.842	--	--
Others	0.41 (0.05–3.53)	0.418	--	--
Chemo during RT	1.36 (0.47–3.91)	0.567	--	--
Cape vs 5FU during RT	0.600 (0.05–6.80)	0.680	--	--
Chemo within 60 days post RT	3.05 (0.35–26.51)	0.312	--	--
Chemo post RT	0.71 (0.30–1.69)	0.441	--	--
Days from diagnosis to RT	1.00 (1.00–1.00)	0.519	--	--
Proton vs photon	0.29 (0.12–0.74)	0.009	0.29 (0.10–0.82)	0.020
RT duration (days)	1.41 (1.06–1.87)	0.018	1.49 (1.07–2.09)	0.020
RT dose (Gy)	0.98 (0.93–1.04)	0.476	--	--
BED (Gy)	0.99 (0.96–1.02)	0.482	--	--
Total GTV (cc)	1.00 (1.00–1.01)	0.150	--	--
Liver volume (cc)	1.00 (1.00–1.00)	0.932	--	--
Mean liver dose (Gy)	1.17 (1.07–1.28)	<0.001	1.19 (1.08–1.32)	<0.001

^*Severe lymphopenia defined as absolute lymphocyte count (ALC) <0.5k/μl within 12 weeks of initiating radiotherapy (RT)

Abbreviations: OR=odds ratio, 95% CI=95% confidence interval, ECOG PS=Eastern Cooperative Oncology Group performance status, WBC=white blood cell, HGB=hemoglobin, PLT=platelet, ANC=absolute neutrophil count, ALC=absolute lymphocyte count, RT=radiotherapy, FOLFOX=5-fluorouracil/oxaliplatin/leucovorin, cape=capecitabine, 5FU=5-fluorauracil, BED=biologically effective dose, GTV=gross target volume

Survival outcomes

Median follow up was 16 months (range 2–85) from the time of RT initiation. At the time of last follow up, at total of 81 patients (84%) had died: 51 (63%) died of disease progression, 6 (7%) died of liver decompensation, 12 (15%) died of biliary infection, and 12 (15%) died of other/unknown causes of death (Table 4). Among patients who developed severe lymphopenia, 55 (85%) had died with cause of death including disease progression (62%, n=34), liver decompensation (7%, n=4), biliary infection (15%, n=8), and other/unknown cause (16%, n=9). Among patients who did not develop severe lymphopenia, 26 (84%) had died with cause of death including metastatic disease (65%, n=17), liver decompensation (8%, n=2), biliary infection (15%, n=4), and other/unknown cause (12%, n=3). There was no significant difference in cause of death between the cohort that developed severe lymphopenia vs not.

Table 4.

Causes of death for 81 patients

Causes, N(%)	All patients (N=81)	No severe lymphopenia* (N=26)	Severe lymphopenia* (N=55)	P value
Disease progression	51 (63)	17 (65)	34 (62)	0.977
Decompensated liver	6 (7)	2 (8)	4 (7)
Biliary infection	12 (15)	4 (15)	8 (15)
Other/unknown	12 (15)	3 (12)	9 (16)

^*Severe lymphopenia defined as absolute lymphocyte count (ALC) <0.5k/μl within 12 weeks of initiating radiotherapy (RT)

The 1-year and 2-year OS rates were 63% and 27% for patients who developed severe lymphopenia vs 77% and 46% for patients who did not develop severe lymphopenia, respectively (Figure 1A; log-rank test p=0.027). When comparing patients who developed vs did not develop Grade 4 lymphopenia, the 1-year and 2-year OS rates were 31% and 21% vs 73% and 36%, respectively (Figure 1B; log-rank test p=0.002).

Figure 1.

Overall survival stratified by (A) severe lymphopenia (absolute lymphocyte count <0.5 k/μl) and (B) Grade 4 lymphopenia (absolute lymphocyte count <0.2 k/μl) within 12 weeks of initiating radiotherapy

On univariable Cox analysis, severe lymphopenia (HR=1.72 [95% CI 1.06–2.79], p=0.029) as well as male sex (HR=1.84 [95% CI 1.18–2.88], p=0.008), higher baseline WBC (HR=1.04 [95% CI 1.01 –1.06], p=0.005), lower baseline albumin (HR=1.72 [95% CI 1.13–2.62], p=0.011), higher baseline CEA (HR=1.05 [95% CI 1.01–1.09], p=0.006), higher baseline CA19–9 (HR=1.00 [95% CI 1.00 –1.00], p=0.036), and biliary obstruction (HR=5.71 [95% CI 3.28–9.95], p<0.001) were significantly associated with increased hazards of death (Table 5). On the multivariable model adjusted for all significant variables from the univariable analysis, severe lymphopenia (HR=1.70 [95% CI 1.02–2.86], p=0.043) as well as biliary obstruction (HR=4.40 [95% CI 2.25–8.60], p<0.001) remained significantly associated with higher hazards of death, while other variables were no longer significant (Table 5). When Grade 4 lymphopenia, which was also significant on univariable Cox analysis (HR=2.66 [95% CI 1.39–5.10], p=0.003), was incorporated to the multivariable model, it remained significantly associated with a greater hazard of death (HR=2.78 [95% CI 1.36–5.67], p=0.005). Other clinical and treatment variables such as age, ECOG performance status, immunosuppression, baseline blood counts including HGB, PLT, ANC, and ALC, number of tumors, tumor size, extrahepatic disease, cirrhosis, macrovascular invasion, receipt of chemotherapy, RT modality, RT dose, cumulative GTV, and MLD were not associated with OS.

Table 5.

Univariable and multivariable Cox regression analysis for time to death from RT

	Univariable		Multivariable
Variables	HR (95% CI)	P value	HR (95% CI)	P value
Age (years)	1.01 (0.99–1.03)	0.348	--	--
Sex, male vs female	1.84 (1.18–2.88)	0.008	1.49 (0.93–2.42)1 / 1.48 (0.91–2.41)2	0.1001 / 0.1152
ECOG PS ≥1 vs 0	1.03 (0.66–1.60)	0.908	--	--
Immunosuppression	1.70 (0.77–3.71)	0.188	--	--
Baseline WBC (k/μl)	1.04 (1.01 –1.06)	0.005	1.05 (1.00–1.10)1 / 1.06 (1.01–1.11)2	0.0741 / 0.0342
Baseline HGB (g/dL)	0.95 (0.83–1.08)	0.420	--	--
Baseline PLT (k/μl)	1.00 (1.00–1.00)	0.658	--	--
Baseline ANC (k/μl)	1.01 (0.96–1.07)	0.674	--	--
Baseline ALC (k/μl)	0.96 (0.65–1.42)	0.828	--	--
Number of tumors present
1	Ref	--	--	--
2	1.22 (0.60–2.48)	0.580	--	--
3	0.75 (0.23–2.40)	0.624	--	--
≥4	1.20 (0.57–2.51)	0.634	--	--
Number of tumors treated
1	Ref	--	--	--
2	0.65 (0.28–1.49)	0.307	--	--
3	0.30 (0.04–2.16)	0.230	--	--
Tumor size (cm)	0.98 (0.91–1.06)	0.632	--	--
Extrahepatic metastasis	1.43 (0.84–2.44)	0.190	--	--
Cirrhosis	1.51 (0.72–3.17)	0.278	--	--
Baseline albumin (g/dL)	0.58 (0.38–0.88)	0.011	0.763 (0.470–1.239)1 / 0.79 (0.50–1.25)2	0.2741 / 0.7932
Baseline INR	1.00 (1.00–1.01)	0.079	--	--
Baseline CEA (ng/ml)	1.05 (1.01–1.09)	0.006	1.01 (0.97–1.06)1 / 1.02 (0.97–1.07) 2	0.6051 / 0.5342
Baseline CA19–9 (U/mL)	1.00 (1.00–1.00)	0.036	1.00 (1.00–1.00)1 / 1.00 (1.00–1.00)2	0.2521 / 0.2742
Biliary obstruction	5.71 (3.28–9.95)	<0.001	4.40 (2.25–8.60)1 / 4.38 (2.23 –8.61)2	<0.0011 /<0.0012
Macrovascular invasion	0.95 (0.58–1.56)	0.851	--	--
Chemo prior to RT	0.71 (0.46–1.10)	0.126	--	--
Chemo during RT	0.74 (0.41–1.32)	0.301	--	--
Chemo post RT	0.85 (0.54–1.34)	0.475	--	--
Days from diagnosis to RT	1.00 (1.00–1.00)	0.055	--	--
Proton vs photon	0.75 (0.47–1.20)	0.227	--	--
RT duration (days)	1.07 (0.98–1.16)	0.134	--	--
RT dose (Gy)	1.01 (0.98–1.04)	0.485	--	--
BED (Gy)	1.01 (0.99–1.02)	0.501	--	--
Total GTV (cc)	1.00 (1.00–1.00)	0.805	--	--
Liver volume (cc)	1.00 (1.00–1.00)	0.780	--	--
Mean liver dose (Gy)	1.02 (0.98–1.06)	0.382	--	--
≥G3 lymphopenia post-RT1	1.72 (1.06–2.79)	0.029	1.70 (1.02–2.86)1	0.043 1
G4 lymphopenia post-RT2	2.66 (1.39–5.10)	0.003	2.78 (1.36–5.67)2	0.005 2

Abbreviations: HR=hazard ratio, 95% CI=95% confidence interval, ECOG PS=Eastern Cooperative Oncology Group performance status, WBC=white blood cell, HGB=hemoglobin, PLT=platelet, ANC=absolute neutrophil count, ALC=absolute lymphocyte count, RT=radiotherapy, BED=biologically effective dose, GTV=gross target volume

Discussion

This is the first study, to our knowledge, evaluating treatment-related lymphopenia and its impact on survival among patients with unresectable ICC. We found that 68% of patients developed severe lymphopenia after initiating HF-RT and had worse OS compared to patients who did not develop severe lymphopenia. While baseline ALC did not correlate with survival, development of severe lymphopenia post-RT predicted worse OS regardless of other prognostic factors including tumor features and liver functions. Our findings suggest that development of severe lymphopenia post-RT is an independent prognostic factor for worse survival in patients undergoing RT for unresectable ICC.

Circulating lymphocytes are the most radiosensitive cells in the blood,¹² with studies demonstrating that relatively low doses of 2–3 Gy are capable of depleting 50–90% of the lymphocyte population.¹³ Prior studies in solid malignancies have suggested that higher RT dose, larger target volume, conventionally fractionated vs hypofractionated RT, and photon vs proton therapy (which limits mean body dose) are predictive of the severity of RT-related lymphopenia.^{22,27,34–36} In this study, we found that receipt of photon vs proton-based RT, higher MLD, and longer duration of RT were predictive of severe lymphopenia. Patients who developed severe lymphopenia also had significantly lower baseline ALC (median 1.1 vs 1.5 k/μl, p=0.005) and larger GTV (median 125 vs 62 cc, p=0.015), although both did not correlate with severe lymphopenia in our logistic regression analysis, perhaps due to lack of statistical power. Other variables such as baseline patient characteristics including immunosuppression, number of tumors treated, baseline labs, liver function tests, presence of cirrhosis, biliary obstruction, receipt of chemotherapy, prescription RT dose, and targeted tumor volume did not correlate with severe lymphopenia in this study.

Compared to other solid tumors, irradiation of hepatic tumors is likely associated with an increased risk of RT-related lymphopenia as the liver is an organ with a rich blood circulation, thus harboring a large pool of circulating lymphocytes that may be exposed to radiation. While there have been no prior studies investigating the prognostic impact of RT-related lymphopenia in ICC, recent studies of liver-directed RT in HCC demonstrated an association between severe lymphopenia and worse survival.^26–29 In a retrospective study of 920 HCC patients by Byun et al., Grade 3 or worse lymphopenia within 3 months of initiating RT was associated with poor OS (HR=1.40, p=0.035), with higher PTV and lower baseline ALC predictive of severe lymphopenia.²⁷ Another study by De et al. including 143 patients with unresectable HCC demonstrated that Grade 3 or worse lymphopenia after RT was associated with worse OS (HR=1.03, p=0.043), with receipt of proton RT and lower body volume of low-dose bath associated with lower risk of RT-related severe lymphopenia.²⁹ In our study of unresectable ICC treated by HF-RT, severe RT-related lymphopenia was associated with approximately 1.7-fold increased risk of death (p=0.045). Moreover, Grade 4 lymphopenia was associated with a 3.1-fold increased risk of death (p=0.002). Similar to prior study findings, patients who developed severe lymphopenia had a significantly larger cumulative GTV and lower baseline ALC, with proton-based therapy predictive of lower likelihood of developing severe lymphopenia.

The association between severe lymphopenia post-RT and worse OS is hypothesized to be related to poor tumor control from compromised immune surveillance. While lymphopenia may be a simple reflection of overall bone marrow suppression which is a poor prognostic factor, the majority of patients in this cohort had normal baseline blood counts and baseline WBC, HGB, PLT, and ANC did not differ significantly between those who developed vs did not develop severe lymphopenia. Furthermore, while infection secondary to lymphopenia may be a significant contributor of death, there was no significant difference with biliary infection as cause of death between patients who developed vs did not develop severe lymphopenia (15% in both). Regardless of the exact mechanism by which RT-related lymphopenia correlates with worse survival, identification of various predictors of lymphopenia as well as management approaches that can minimize the severity of lymphopenia is important. Based on our findings, potential approaches include closer observation in high-risk patients (i.e., lower ALC at baseline, larger tumor volume), minimizing MLD, use of proton-based RT, and avoiding treatment breaks that prolong total duration of RT.

Our findings highlight the importance of immune surveillance in the survival outcomes of ICC and indirectly lends further support of the role of immunotherapy in the treatment of ICC. In the KEYNOTE-028 phase 1b and KEYNOTE-158 phase 2 prospective studies investigating the safety and efficacy of pembrolizumab, a monoclonal anti-PD-1 antibody, immune checkpoint blockade (ICB) has shown promising results in advanced biliary tract cancers.^37,38 In the recent KEYNOTE-966 and TOPAZ-1 phase 3 trials, pembrolizumab and durvalumab (PD-L1 inhibitor) in addition to chemotherapy, respectively, have been shown to significantly improve OS in patients with advanced biliary tract cancer, establishing the addition of immunotherapy as standard of care.^6,39 With the increasing use of immunotherapy for advanced ICC as well as data supporting association between lymphopenia and reduced benefit from ICB,^40–43 minimization of lymphopenia in patients undergoing local therapy with RT will be important in optimizing response to ICBs.

The present study has several limitations including its retrospective nature, small sample size, and heterogenous patient and treatment characteristics. Given this study was retrospective, patients had blood counts available at various time points since initiating RT and our primary variable of interest, severe lymphopenia, represents a relatively wide time range of 12 weeks, though majority of patients (75%) had ALC nadir within 30 days of starting RT. The significance of ALC at different time points relative to RT remains to be further explored. Additionally, our study cohort represents a heterogeneous group of patients who underwent various treatment regimens. For instance, 52% underwent chemotherapy prior to RT, 7% had concurrent chemotherapy, and 40% had chemotherapy post-RT with varying chemotherapy regimens. The impact of different chemotherapy timing and agents on lymphopenia and survival remains unclear, though receipt of chemotherapy at various times with respect to RT did not correlate with severe lymphopenia in our analysis. Of note, only five (5%) patients received immunotherapy (durvalumab with gemcitabine/cisplatin) in this study cohort, limiting the evaluation of the effect of lymphopenia in patients of undergoing immunotherapy. Lastly, this study includes a small sample size of 96 patients which limits the strength of our analysis. Despite these limitations, we found a significant association between severe lymphopenia post RT and survival and our results are hypothesis-generating. Larger, prospective studies addressing the limitations of this study will be important in further establishing the relationship between RT-related lymphopenia and clinical outcomes.

In conclusion, more than half of patients undergoing HF-RT for unresectable ICC develop severe lymphopenia which is a potential prognostic factor for worse survival. The association between severe lymphopenia post-RT and survival observed in this study supports the important role of the host immune system in the outcomes of ICC. With increasing role and use of immunotherapy for patients with advanced ICC, minimizing RT-related lymphopenia in patients undergoing local therapy may help optimize response to ICB. Patients with lower baseline ALC and larger tumor volume may be at increased risk, and use of proton-based RT, minimizing MLD, and avoiding breaks in RT that prolong total duration of RT may be strategies to mitigate the risk of developing RT-related severe lymphopenia. Large, prospective studies are needed to further establish the relationship between RT-related lymphopenia, its predictors, and outcomes in unresectable ICC.