Oncologic outcomes of intraoperative autologous blood transfusion for major oncologic resection

Abstract

Background:

Intraoperative autologous blood transfusion (IABT) is contraindicated during oncologic resection due to concern for metastasis. However, there is a paucity of data to substantiate this claim, and the true risk of IABT in patients with cancer remains unknown.

Purpose:

We identified patients who underwent oncologic resection with IABT 2010-2021 at our institution. The primary outcome was early metastasis. A subgroup analysis of patients who underwent liver transplant for malignancy was conducted using inverse probability of treatment weighting to compare survival between those who received autologous and allogeneic transfusions.

Results:

Of 444 patients, the most common diagnoses were hepatocellular carcinoma (235, 52.9%), cholangiocarcinoma (68, 15.3%), and renal cell carcinoma (18, 4.1%). The median volume of autologous blood transfused was 661 ml (IQR 337-1491 ml). Seven patients (1.6%) experienced distant recurrence within 90 days, and 35 (7.9%) within one year. Of the seven patients with early distant recurrence, only one had metastasis not attributable to preoperative factors. In a subgroup analysis of patients undergoing liver transplant for malignancy, 299 patients who received IABT were compared to 107 patients who received allogeneic transfusion. After adjusting for cancer type, age, sex, estimated blood loss, MELD score, stage, and allograft type, IABT was not associated with OS [AHR 1.30 (95% CI 0.82-2.03), p=0.241] or RFS [AHR 1.15 (95% CI 0.77-1.73), p=0.498].

Conclusion:

IABT does not appear to be associated with early recurrence in patients undergoing oncologic resection. IABT should be investigated as a possible alternative to massive transfusion of allogeneic blood during oncologic resection.

Introduction

Radical oncologic resection of malignant solid tumors often involves the potential for significant blood loss and the need for blood transfusion. [1, 2] Allogeneic blood transfusion has been associated with an increased risk of postoperative complications [3-7], an increased risk of cancer recurrence, and increased mortality in cancer patients. [8-16] Increasing recognition of these deleterious effects has prompted a search for alternative strategies to minimize the frequency and volume of allogeneic transfusions. [17-19] In situations where transfusion cannot be avoided, intraoperative autologous blood transfusion (IABT) has emerged as a promising alternative to mitigate potential complications associated with allogeneic transfusion in non-cancer operations. [20]

The use of IABT in cancer surgery has traditionally been avoided due to concerns regarding the potential development of early distant metastasis from the reintroduction of cancer cells in salvaged blood. This concern was first identified in a case report from 1975 in which the blood of a patient undergoing pneumonectomy for non-small cell lung cancer was found to contain malignant cells after passing through an autotransfusion filter. [21] Several subsequent studies confirmed the presence of malignant cells in salvaged blood. [22, 23] Although it was recognized that these malignant cells suffered mechanical trauma during the salvage process and their viability was never established, the presence of these cells led to the assumption that the use of IABT would disseminate tumor cells throughout the bloodstream and result in distant metastasis. As a result, malignancy is considered a contraindication to IABT.

Over the past several decades the biologic mechanisms of metastasis have been elucidated and the dogma discouraging the use of IABT in cancer patients has been questioned. It is increasingly recognized that circulating tumor cells are present in the blood of many cancer patients regardless of stage, and that seeding of distant metastasis involves a complex interplay between malignant cells, receptive tissue, and the immune system. [24] These discoveries have prompted reconsideration of blood salvage in cancer surgery. Studies on liver transplant for hepatocellular carcinoma have consistently found no association between the use of IABT and cancer recurrence. [25-31] Although several studies on other types of malignancy appear to support this finding, they primarily consist of small series or lack data on oncologic outcomes. [32-37] No study has yet assessed the oncologic safety of IABT across a spectrum of cancer types.

In this study, we investigated short-term outcomes of radical oncologic resections of a variety of solid tumor malignancies in which IABT was used to determine if the theoretical concern for the rapid development of widespread metastasis is reflected in empirical data.

Methods

Study cohort

The Mayo Clinic Institutional Review Board approved this study. We retrospectively identified patients who underwent resection of a solid tumor malignancy at Mayo Clinic in Rochester, Minnesota between 2010 and 2021 in which intraoperative autologous blood transfusion was used. Patients who were under 18 years old, underwent aborted or incomplete resection, required cardiopulmonary bypass, or were lost to follow-up within 90 days of surgery were excluded from the study.

Intraoperative autologous transfusion protocol

Intraoperative autologous transfusion was used at the discretion of the surgeon and anesthesiologist, typically in the setting of significant life-threatening bleeding for non-transplant oncologic resections. Use of intraoperative autologous transfusion is standard in liver transplant cases, but not all cases had adequate blood loss for salvaged blood to be transfused. Cell salvage was performed using the autoLog Legacy autotransfusion system (Medtronic, Minneapolis, MN). The patient’s blood was suctioned from the operative field to the device reservoir, and the volume of collected blood was recorded for accurate estimation of blood loss. When the volume of whole blood reached 350 ml the blood was filtered, centrifuged, and washed with saline. The washed red blood cells were then transfused into the patient. A leukocyte depletion filter was used in all cases after March 2012. [38]

Data collection

Clinicodemographic data were collected from review of medical record. Pathology reports were used to verify and classify type of malignancy. Transfusion volumes were identified from an internal blood bank database. Date of recurrence was defined as the date of radiographic suspicion of recurrence that was subsequently confirmed by cytologic or histologic evaluation.

Statistical analysis

Categorical data are expressed as counts and proportions, and continuous data are expressed as median and interquartile range (IQR) unless otherwise stated.

Entire cohort

The primary outcome for the entire cohort was distant recurrence at 90 days after surgery. Patients without the outcome of interest were required to have at least 90 days of follow-up to be included in this cohort. Surgical and oncologic details of each patient with this outcome were further assessed and reported individually. Inferential statistics were not performed for this cohort given patient heterogeneity.

Subgroup analysis

After identifying a comparison cohort of patients who underwent liver transplant for malignancy with allogeneic red blood cell transfusion, subgroup analysis was conducted to compare oncologic outcomes between those who received autologous and allogeneic transfusions. Differences between the autologous and allogeneic groups were compared using Wilcoxon Rank Sum tests for continuous variables, and Chi-square or Fisher’s exact tests for categorical variables as appropriate. Predicted probabilities of undergoing autologous transfusion were estimated after adjusting for cancer type, sex, age at time of surgery, estimated blood loss (EBL), end stage liver disease (MELD) score, tumor stage, nodal stage, and allograft type (living or deceased donor). Overall survival (OS) and recurrence free survival (RFS) were evaluated. OS was defined as the time from surgery to the date of death or last follow-up. RFS was defined as the time from surgery to recurrence, death, or last follow-up. Factors associated with OS and RFS were assessed with inverse probability of treatment weighting (IPTW)-adjusted multivariable Cox proportional hazards models. A two-sided p-value of <0.05 was considered significant. Analysis was performed with R (Version 4.3.1, R Foundation for Statistical Computing, Vienna, Austria) and SAS (Version 9.4, SAS Institute Inc., Cary, NC).

Results

A total of 715 patients with a pathologically confirmed solid tumor malignancy underwent an operation in which autologous transfusion was used between 2010 and 2021. One hundred ninety-one patients were excluded from the study because they received autologous transfusion in an operation other than oncologic resection, 66 patients underwent aborted or incomplete resection, and 14 patients were lost to follow-up before 90 days. The remaining 444 patients were included in our analysis.

Cohort characteristics are listed in Table 1. Three hundred ten patients (69.8%) were male and 134 patients (30.2%) were female. The median age was 59.5 years (IQR 53.9-65.0 years). The most common cancer types were hepatobiliary (325 patients, 73.2%), urologic (29 patients, 6.5%), and thoracic (excluding esophagus; 22 patients, 5.0%). Specialties that most often employed IABT in oncologic resection were transplant surgery (311, 70%), thoracic surgery (35 patients, 7.9%), and urology (29 patients, 6.5%). Among the 259 patients for which information on estimated blood loss was available, the median volume of estimated blood loss during the procedure was 1,530 ml (IQR 650-3,108 ml). The median volume of autologous blood transfused among all patients in the cohort was 661 ml (IQR 337-1491 ml). Eleven patients (2.5%) died within 90 days of surgery. Six of these deaths were due to postoperative complications, and five were due to allograft failure after liver transplant. None were attributable to metastatic disease.

Table 1:

Cohort characteristics

	n = 444
Age	59.5 (53.9-65.0)
Sex
Male	310 (69.8%)
Female	134 (30.2%)
Cancer type
Hepatobiliary	325 (73.2%)
Hepatocellular carcinoma	235 (52.9%)
Cholangiocarcinoma	68 (15.3%)
Liver (other)	10 (2.3%)
Pancreatic ductal adenocarcinoma	7 (1.6%)
Pancreatic neuroendocrine tumor/carcinoma	4 (1.0%)
Gallbladder cancer	1 (0.2%)
Urologic	29 (6.5%)
Renal cell carcinoma	18 (4.1%)
Bladder cancer	6 (1.4%)
Testicular cancer	3 (0.7%)
Prostate cancer	2 (0.5%)
Thoracic (excluding esophagus)	22 (5.0%)
Non-small cell lung cancer	13 (2.9%)
Mediastinal malignancy	9 (2.0%)
Gynecologic	20 (4.5%)
Uterine cancer	10 (2.3%)
Ovarian cancer	8 (1.8%)
Gynecologic malignancy (other)	2 (0.5%)
Gastrointestinal	19 (4.3%)
Esophageal cancer	12 (2.7%)
Colorectal cancer	6 (1.4%)
Gastric cancer	1 (0.2%)
Sarcoma	14 (3.2%)
Extremity sarcoma	8 (1.8%)
Retroperitoneal sarcoma	6 (1.4%)
Brain malignancy	8 (1.8%)
Head & neck cancer	4 (1.0%)
Other	3 (0.7%)
Surgical service
Transplant	311 (70.0%)
Thoracic surgery	35 (7.9%)
Urology	29 (6.5%)
Gynecologic oncology	20 (4.5%)
Hepatobiliary surgery	17 (3.8%)
Neurosurgery	8 (1.8%)
Orthopedic surgery	8 (1.8%)
Colorectal surgery	5 (1.1%)
Otolaryngology	5 (1.1%)
Vascular surgery	4 (1.0%)
Cardiovascular surgery	1 (0.2%)
Breast & melanoma surgical oncology	1 (0.2%)
Estimated blood loss	1,530 ml (650-3,108 ml)
Autologous transfusion	444 (100%)
Volume autologous transfusion	661 ml (337-1491 ml)
90-day mortality	11 (2.5%)

Continuous variables: median (IQR)

Categorical variables: count (percent)

Seven patients (1.6%) experienced distant recurrence within 90 days of surgery. Out of the seven patients with distant recurrence within 90 days, all but one patient had preoperative factors that explained the early recurrence. This included two patients who had preoperative metastasis that progressed, two patients whose final pathology result in significant upstaging of the cancer, one patient with metastasis that was identified intraoperatively, and one patient with tumor marker levels that in retrospect had prohibitively high risk of early recurrence. Further details regarding these recurrences are included in Table 2. At one year after surgery 35 patients (7.9%) experienced distant recurrence.

Table 2:

Patients with distant recurrence within 90 days of surgery

	Operation	Pathology	Days to recurrence	Location of recurrence	Notes
1	Liver transplant	HCC1	20	Retroperitoneal LN Bone	Grossly exceeded Milan criteria Would have precluded surgery
2	Cystoprostatectomy & left lateral sectorectomy	Urothelial carcinoma	39	Liver	Found metastatic at time of resection Would have precluded surgery
3	En bloc multi-visceral resection	PDAC2	78	Pancreatic tail Liver	CA 19-9 11,500 two days prior to surgery Multiple liver metastases found 78 days after surgery
4	Nephrectomy	RCC3	82	Liver Lung	Lung nodule present on preoperative imaging
5	Nephrectomy	RCC3	84	Lung	Two lung nodules discovered 84 days after surgery
6	Liver transplant	CCA4	86	Liver Lung	Thought to be HCC prior to transplant Found 6 separate lesions of CCA on final pathology, 0.3 cm to 7 cm Would have precluded surgery
7	Nephrectomy	RCC3	90	Lung	Lung nodule present on preoperative imaging

¹Hepatocellular carcinoma

²Pancreatic ductal adenocarcinoma

³Renal cell carcinoma

⁴Cholangiocarcinoma

Subgroup analysis

In a subgroup analysis of patients who underwent liver transplant for malignancy, 107 patients who received allogeneic transfusion were identified as a control group for comparison with 299 patients who received autologous transfusion. Subgroup characteristics are shown in Table 3. Patients who received autologous transfusion were more likely to have hepatocellular carcinoma, higher volumes of blood loss, higher MELD score, and a deceased donor allograft. Unadjusted Kaplan-Meier curves separated by cancer type are shown in Figure 1. Log rank tests were not performed prior to adjustment. After adjusting for propensity score weight with the IPTW method, factors associated with both OS and RFS were cancer type, sex, and tumor stage, as shown in Table 4. Use of autologous transfusion was not associated with OS (HR 1.30, p=0.241) or RFS (HR 1.15, p=0.498).

Table 3:

Cohort characteristics of patients who underwent liver transplant for malignancy

	Allogeneic transfusion (n = 107)	Autologous transfusion (n = 299)	Total (n = 406)	p-value
Diagnosis
HCC1	35 (32.7%)	234 (78.3%)	269 (66.3%)	<0.001
CCA2	72 (67.3%)	65 (21.7%)	137 (33.7%)
Sex
Female	32 (29.9%)	66 (22.1%)	98 (24.1%)	0.10
Male	75 (70.1%)	233 (77.9%)	308 (75.9%)
Age	56.8 (46.2-63.9)	59.1 (53.6-63.1)	58.7 (50.5-63.6)	0.11
EBL3
≤2000 ml	88 (82.2%)	167 (55.9%)	255 (62.8%)	<0.001
>2000 ml	15 (14.0%)	127 (42.5%)	142 (35.0%)
Missing	4 (3.7%)	5 (1.7%)	9 (2.2%)
MELD4	9.0 (7.0-15.0)	13.0 (9.0-21.0)	12.0 (9.0-19.0)	<0.001
Tumor stage
T1	71 (66.4%)	148 (49.5%)	219 (53.9%)	0.008
T2	31 (29.0%)	137 (45.8%)	168 (41.4%)
T3	5 (4.7%)	14 (4.7%)	19 (4.7%)
Nodal stage
N0	99 (92.5%)	297 (99.3%)	396 (97.5%)	0.001
N1	8 (7.5%)	2 (0.7%)	10 (2.5%)
Metastasis
M0	106 (99.1%)	299 (100.0%)	405 (99.8%)	0.26
M1	1 (0.9%)	0 (0.0%)	1 (0.2%)
Allograft type
Deceased donor	76 (71.0%)	274 (91.6%)	350 (86.2%)	<0.001
Living donor	31 (29.0%)	25 (8.4%)	56 (13.8%)
HCC patients
(n = 269)
Milan criteria met	12 (34.3%)	87 (37.2%)	99 (36.8%)
No	23 (65.7%)	146 (62.4%)	169 (62.8%)	0.87
Yes	0 (0.0%)	1 (0.4%)	1 (0.4%)
NA
	8.8 (3.7-10.0)	4.9 (2.8-12.0)	5.1 (3.0-13.0)	0.038
AFP5
CCA patients
(n = 137)	63.5 (19.0-150.0)	50.0 (13.0-149.0)	56.0 (18.0-150.0)	0.18
CA 19-9

Continuous variables expressed as median (IQR)

¹Hepatocellular carcinoma

²Cholangiocarcinoma

³Estimated blood loss

⁴Model for End-Stage Liver Disease

⁵Alpha Fetoprotein

Figure 1:

Unadjusted Kaplan-Meier survival curves for patients who underwent liver transplant for malignancy

Table 5:

Multivariable Cox proportional hazards model adjusted for propensity score weight through inverse probability of treatment weighting method

	Overall Survival		Recurrence-free Survival
	Hazard Ratio (95% confidence interval)	p-value	Hazard Ratio (95% confidence interval)	p-value
Autologous transfusion	1.30 (0.84-2.03)	0.24	1.15 (0.77-1.73)	0.50
Diagnosis (CCA1 vs HCC2)	1.86 (1.02-3.37)	0.042	1.93 (1.13-3.31)	0.016
Age at surgery (per 10 years)	1.01 (0.82-1.25)	0.94	1.03 (0.85-1.25)	0.79
Sex (male vs female)	0.54 (0.33-0.87)	0.012	0.48 (0.31-0.74)	0.001
EBL3 category (≤2000 ml vs >2000 ml)	0.96 (0.58-1.61)	0.90	0.77 (0.48-1.23)	0.27
MELD4 (per 5 units)	1.00 (0.90-1.11)	0.96	0.98 (0.88-1.10)	0.74
Tumor stage
T2 vs T1	1.01 (0.59-1.73)	0.96	1.01 (0.61-1.69)	0.96
T3 vs T1	5.95 (3.39-10.43)	<0.001	5.95 (3.38-10.46)	<0.001
Nodal stage (N1 vs N0)	1.24 (0.41-3.77)	0.70	1.31 (0.37-4.66)	0.68
Allograft type (living vs deceased donor)	1.63 (0.86-3.08)	0.13	1.27 (0.67-2.38)	0.47

¹Cholangiocarcinoma

²Hepatocellular carcinoma

³Estimated blood loss

⁴Model for End-Stage Liver Disease

Discussion

Although current use of IABT in cancer surgery is limited due to concern for tumor seeding, we found that early widespread distant metastasis after IABT during oncologic resection is rare. This finding was consistent throughout a heterogenous cohort with a variety of cancer types, stages, and operations. Among patients undergoing liver transplant for malignancy of any type, use of IABT did not appear to be associated with overall survival or recurrence-free survival.

This is, to our knowledge, the largest cohort of patients receiving autologous transfusion during oncologic resection that has been reported to date. These results augment the growing body of evidence establishing autologous transfusion as a safe alternative to allogeneic transfusion during cancer surgery. Despite early concerns, contemporary studies across a range of solid tumor malignancies have not found an association between IABT and recurrence. [28, 33, 35, 36, 39-41] Several series, in fact, have demonstrated improved overall and recurrence-free survival with the use of autologous blood when compared with allogeneic transfusion. [32, 42, 43] Despite these findings there has not been widespread adoption of IABT for oncologic resections, and standard practice in the setting of massive intraoperative hemorrhage remains massive transfusion of allogeneic blood.

The oncologic consequences of allogeneic transfusion are well-documented across a spectrum of cancer types, stages, and treatment settings. Receipt of allogeneic blood has been identified as an independent predictor of lack of response to immunotherapy in patients with metastatic cancer [44] and worse OS and RFS in patients receiving definitive chemoradiation for head and neck cancer [45]. In patients undergoing curative-intent surgery for solid tumor malignancies, as in our cohort, intraoperative allogeneic transfusion has been associated with worse OS and RFS in esophageal cancer [9], gastric cancer [16], gallbladder cancer [12], extrahepatic biliary malignancies [13], colorectal cancer [11], and adrenocortical carcinoma [10]. In a study of patients undergoing resection of non-small cell lung cancer, the volume of transfused allogeneic blood was associated with decreased OS and RFS in a dose-dependent manner. [14]

A prominent advantage of autologous transfusion is avoidance of the immune challenge and alloantibody development inherent to allogeneic transfusion. The immunosuppressive effects of allogeneic blood transfusion were first described in renal transplant recipients when it was found that patients who received more blood transfusions were less likely to experience allograft rejection. [46] While dampening immunity is beneficial in transplant patients, transfusion-related immunomodulation is thought to be the mechanism underlying the association between allogeneic transfusion and tumor recurrence in cancer patients. [47] A second well-described immune response to allogeneic transfusion is the development of alloantibodies, which renders each subsequent transfusion more challenging to crossmatch and places the recipient at increased risk of incompatibility reactions. This is of particular relevance to patients with cancer who often require multiple transfusions throughout their treatment. Autologous transfusions obviate both of these risks; avoiding the use of IABT in cancer operations due to unsubstantiated concerns of metastasis denies the advantages of autologous transfusion to a population that could most stand to benefit from them.

Our institution performs liver transplant for both hepatocellular carcinoma and cholangiocarcinoma in select patients, in accordance with Organ Procurement and Transplantation Network transplant oncology guidelines. [48] As liver transplants often cause large-volume blood loss, transfusion of either allogeneic or autologous nature is common, which provided an opportunity for a subgroup analysis of this population. While IABT during liver transplantation for cholangiocarcinoma has been previously described, this operation is less commonly performed for cholangiocarcinoma which is a unique strength of our study.

Our study has several limitations in addition to those inherent to retrospective reviews. Given the current reluctance to use salvaged blood in cancer surgery, at our institution IABT is typically used only in the setting of life-threatening bleeding. Our cohort thus represents a small fraction of those undergoing oncologic resection, which may limit generalizability of results. However the operative conditions that would prompt the decision to employ IABT despite existing concerns, such as profound anemia or excessive bleeding, are conditions that have been independently associated with operative mortality and early recurrence. [17, 49-53] We would therefore expect that our cohort would have greater than average risk of poor short-term outcomes; that this was not seen in our study further supports the safety of autologous transfusion.

Similarly, the heterogeneity of our cohort is both a limitation and a strength of our study. Our cohort is comprised of patients with a wide range of cancer types of all stages undergoing a variety of resections. While this heterogeneity enhances the generalizability of our main finding that the use of autotransfusion does not result in early widespread metastasis, it prevented us from performing a robust subgroup analysis of the non-transplant patients. Further studies are needed to investigate possible distinctions between the types of malignancies that are analyzed in aggregate in this study. Our findings suggest that these studies are safe to pursue.

Conclusion

Perioperative blood transfusion in patients undergoing oncologic resection is common, and using autologous blood in these patients offers substantial advantages over allogeneic blood. The long-held concern that autologous transfusion will result in early widespread metastasis is not supported by our investigation. These data suggest the use of IABT could be considered in the setting of life-threatening bleeding during oncologic resection and support the routine use of IABT in transplant patients. Our findings underscore the need for additional translational studies and the development of a randomized controlled trial for solid tumor resection.

Synopsis:

In patients undergoing oncologic resection, we found no evidence that use of intraoperative autologous transfusion caused early metastasis. There was no difference in OS or RFS among patients undergoing liver transplant for malignancy between those with autologous vs. allogeneic transfusion.