Abstract
Introduction:
Return to intended oncologic treatment (RIOT) is a patient-centric quality metric that has been associated with survival in multiple malignancies. Emergency general surgery conditions, such as cholecystitis, are frequent events that can disrupt treatment.
Methods:
Our institutional database was queried using ICD 9/10 codes for patients on active cancer therapy who developed cholecystitis between 01/01/2010 and 12/31/2020. The primary outcome was RIOT ≤ 6-weeks stratified by whether patients were treated with surgery, drainage procedure, or no procedural management.
Results:
Out of 129 patients, 52.7% (N=68) returned to therapy within 6-weeks whereas (22.5%) (N=29) had a >6-week delay and 24.8% (N=32) never received additional cancer-directed therapy. RIOT ≤ 6-weeks was associated with lower ECOG scores (ECOG 0–1 81.8% vs. 52.5%, p=0.023) and patients were more likely to have solid (57.4%) or multiple malignancy types (23.5%). Patients with hematologic malignancies (OR 0.41, 95% CI 0.17–0.98, P=0.044) and those who underwent drainage procedures (OR 0.32, 95% CI 0.13–0.81, P=0.016) were less likely to RIOT ≤ 6 weeks. Only 3 patients who had percutaneous cholecystotomy tube (PCT) placed (8.9%) would later undergo cholecystectomy.
Conclusions:
Half of patients who develop cholecystitis while on cancer treatment RIOT ≤ 6-weeks, and one-quarter will never return to treatment. Patients with hematological malignancies and those who have drainage procedures are less likely to RIOT ≤ 6-weeks. There is a low rate of interval cholecystectomy after (PCT) placement and PCT is unlikely to expedite a return to cancer-directed therapy.
Article Summary
This was retrospective analysis of 129 patients with cancer on active treatment who developed cholecystitis and were managed with surgery, drainage, or no procedural intervention. 52.7% of patients never return to their intended oncologic therapy within 6-weeks, 24.8% never return at all, and gallbladder drainage with associated the a 7.3-fold higher odds of 90-day mortality.
Graphical Abstract
Introduction
Advances in cancer therapy have transformed many malignancies into chronic, treatable diseases. Continuity of oncologic therapy is associated with long-term outcomes, and interruptions in cancer-directed treatment may compromise both disease control and survival1,2. Common reasons for treatment delay or interruption include chemotherapy-related adverse events and surgical complications, however intercurrent emergency general surgery conditions are also frequent occurrences3. In an analysis using the SEER-Medicare database, our group found that patients with breast, lung, and prostate cancer had a 4.8% incidence of an emergency general surgery (EGS) condition within the first year of diagnosis, and a 10-year cumulative incidence of 17%3. In this context, the concept of return to intended oncologic treatment (RIOT) is a pragmatic and patient-centered metric for assessing how interventions for EGS conditions can restore cancer treatment continuity and improve patient outcomes. RIOT was originally described by Aloia et al in 2014, to describe return to cancer therapy after hepatectomy for metachronous metastatic colorectal cancer but has since been researched in multiple gastrointestinal malignancies4,5. Although this topic has been less studied in hematologic malignancies, there are existing data that amount of chemotherapy received and adherence to cancer-directed treatment are associated with improved survival6,7
EGS conditions are a particularly disruptive class of events in the general population, representing 11% of surgical admissions and 50% of surgical mortality8. There is limited data on their impact on cancer patients actively receiving systemic therapy. Several studies have demonstrated worse outcomes among cancer patients undergoing emergency surgery9–11. Cholecystitis is of particular interest because it can range in severity and can be managed with cholecystectomy, a drainage procedure, such as percutaneous cholecystostomy tube (PCT) or endoscopic cholecystoduodenostomy, or treated with antibiotics. Cancer patients are more likely to receive non-operative management compared to non-cancer patients, suggesting either provider bias or patient factors that dictate treatment3,12. Anecdotally, oncology patients presenting with biliary pathology are often treated non-surgically with the expectation it will allow patients a faster return to therapy, however this has never been demonstrated, and it is unlikely there will ever be a clinical trial due to the heterogeneity of the patient population, the wide-ranging severity of cholecystitis at presentation, and a lack of clinical equipoise.
The primary objective of this study was to describe the frequency with which cancer patients on active systemic therapy who develop acute cholecystitis RIOT within 6-weeks, and whether this differs by treatment strategy (surgery, drainage, or non-procedural management). We evaluated 90-day mortality as a secondary outcome to determine if there was an association between treatment strategy and survival.
Methods
Cohort Creation
After receiving Institutional Review Board approval, we queried our institutional database (IDB), for patients who presented to the urgent care center (UCC) with biliary-related EGS conditions between 01/01/2010 and 12/31/2020. Visits were identified using International Classification of Diseases (ICD)-9 and -10 codes for cholecystitis, choledocholithiasis, cholangitis, biliary tract obstruction, and “other,” which included hydrops, perforation, gangrenous cholecystitis, and gallbladder fistula (ICD-9 574.xx, 575.xx, 576.xx and ICD-10 K80.xx, K81.x, K82.x, Supplemental Table 1 and 2). If a patient had multiple encounters for the same condition, the first was used for analyses. Demographics were extracted from the electronic medical record (EMR) and comorbidities were determined by ICD-9 codes and grouped as: chronic liver disease, chronic obstructive pulmonary disease (COPD), diabetes, heart disease, and kidney disease (Supplemental Table 3). Primary malignancies were identified using ICD-10 clinical modification (ICD-10-CM) codes and grouped as hematologic, solid (breast, colorectal, gynecologic, lung, upper gastrointestinal, urologic), and “multiple” in the case of multiple primary malignancies. During cohort creation we excluded patients with malignancies involving the hepatopancreatobiliary system to eliminate confounding by malignancy-induced cholecystitis, cholangitis, and biliary tract obstruction (duodenal C170, liver NOS C220, intrahepatic bile duct C221, gallbladder C239, extrahepatic bile duct C240, ampulla of vater C241, and pancreas C250–252, C254, C257–259). Procedural and operative intervention was identified based on ICD-9-CM, ICD-10-Procedure Coding System (ICD-10-PCS) and Current Procedural Terminology (CPT) codes that pertained to cholecystectomy (open or laparoscopic), PCT placement, and endoscopic internal drainage (e.g., cholecystoduodenostomy) (Supplemental Table 4). Patients were then manually reviewed to identify true cholecystitis and exclude cases of choledocholithiasis, gallstone pancreatitis, and malignant biliary tract obstruction. Patients who had not received oncologic therapy in the 6-weeks prior to UCC admission were then excluded. Patients were then stratified into three groups by treatment: surgery, drainage (which included PCT and internal drainage), or no procedural intervention (which included patients treated with antibiotics alone). Patients who had a PCT placed and subsequently went on to have surgery were included in the surgery cohort (N=3).
Primary and Secondary Outcomes
The primary outcome of the study was return to intended oncologic therapy (RIOT) within (≤) 6-weeks of UCC presentation. Oncologic therapy included cytotoxic chemotherapy, targeted therapy with monoclonal antibodies, immunotherapy, investigational medications, and anti-hormonal therapy in the cases of hormone-receptor positive breast cancer and castration therapy for prostate cancer. Six weeks were chosen both as an intuitive and clinically meaningful period of convalescence after EGS event13, but also because 42 days was the median time from open surgery to RIOT in the original manuscript by Aloia et al14. In this study, we use it to describe returning back to the same therapy a patient was on prior to the interruption. Our secondary outcome was 90-day mortality stratified by treatment arm (surgery, drainage procedure, no procedure).
Statistical Analysis
Categorical variables were summarized using count and percentage and compared using Fisher’s exact test. Continuous variables were summarized using median and IQR and compared using Wilcoxon rank sum test. Logistic regression was used to test associations between clinical and demographic variables and both RIOT within 6 weeks and 90-day mortality. A multivariable model for RIOT ≤ 6-weeks was built using variables with a P-value of 0.2 or less in the univariable analysis. ECOG (Eastern Cooperative Oncology Group) Score was excluded due to a high rate of missingness, which created model instability. We assessed for association between ECOG score, drainage procedures, and surgery were Fisher’s exact test. SAS 9.4 was used for all analysis. All tests were two-sided and P < 0.05 was considered significant.
Results
Initial IDB extraction identified 483 patients, of whom 290 had received chemotherapy 6-weeks prior to UCC presentation. Of these, 161 were excluded for the following reasons: another medication was flagged as chemotherapy (N=4), cholecystitis was in the differential but was not confirmed (N=18), patient was extracted due to prior history of cholecystitis or cholecystectomy (N=31), and “other” (N=47), which included drug-related or sepsis-related transaminitis or hyperbilirubinemia, liver abscess, progression of disease, and for many we could not determine why they were extracted by the IDB request. Based on the inclusion criteria above, 61 patients had a diagnosis other than cholecystitis: cholelithiasis (N=10), choledocholithiasis and gallstone pancreatitis (N=22), and despite the exclusion criteria, 29 had a malignant biliary tract obstruction (Figure 1). This left a final analytic cohort of 129 patients. When stratified by treatment type, 31 patients had a destination PCT (24.0%), 1 had internal endoscopic drainage (0.8%), 35 had surgery (27.1%) and 62 (48.1%) had no procedural intervention. Three patients who had a prior PCT placed went on to have surgery, thus, 34 PCTs were placed in total for a total conversion rate of 8.9%. The 35 surgical patients included 16 (45.7%) who had a minimally invasive (laparoscopic or robotic) procedure, 19 (54.3%) who underwent an open cholecystectomy. Of those who had an open cholecystectomy, 5 (26.3%) were planned as an open approach a priori due to factors such as small bowel volvulized around a PCT, concomitant bowel perforation, and free gallbladder perforation. The remaining 14 (73.7%) were converted from a laparoscopic approach due to adhesions, inflammation around the gallbladder or porta hepatis, or an inability to dissect the critical view of safety.
Figure 1.
Flowchart and Cohort Selection
When stratifying by time to return to oncologic therapy, 52.7% (N=68) returned in ≤6-weeks, whereas 47.3% (N=61) did not, which included 32 patients (24.8% of full cohort) who never returned to cancer-directed therapy at all (Table 1). Among the 32 who never returned to therapy, 12 died during the hospitalization associated with their EGS event (37.5%) and 14 were medically unfit for surgery (43.8%), which included 8 patients who transitioned to hospice or were made DNR during that hospitalization. The remainder either couldn’t tolerate or declined to restart therapy. Five (15.6%) underwent cholecystectomy, while the remainder were treated with no procedure or drainage (no procedure N=11 [34.4%], drainage N=16 [50%]). None of the 5 patients who underwent surgery but failed to return to therapy died in the hospital. Patients who returned to intended therapy within 6-weeks had lower ECOG scores (ECOG 0–1 81.8% vs. 52.5%, p=0.023) and were more likely to have solid (57.4% vs. 49.2%) or multiple malignancies (23.5% vs. 13.11%). Patients who did not RIOT within 6-weeks were more likely to have hematologic malignancies (37.7% vs. 19.1%) (P=0.044 for all). A higher proportion of patients who returned to therapy in ≤ 6-weeks underwent surgery (32.4% vs. 21.3%) compared to drainage (14.7% vs. 36.1%, P=0.018). Moreover, 62.9% of patients who underwent surgery returned to therapy in ≤ 6-weeks, compared to 32.3% of patients who received a drainage procedure There was no statistically significant association between ECOG score, drainage procedures, and surgery (P=0.718, Supplemental Table 5).
Table 1.
Demographics
| Variable | RIOT ≤ 6-weeks (N=68) | Late or No RIOT (N=61) | Total (N=129) | P-value |
|---|---|---|---|---|
| Age (median, IQR) | 59.8 (60.6, 69.6) | 61.5 (53.8, 73.4) | 61 (52.9, 70.2) | 0.27 |
| Female Gender (N, %) | 39 (57.4%) | 36 (59.0%) | 75 (58.1%) | 0.86 |
| White Race (N, %) | 53 (79.1%) | 44 (73.3%) | 97 (76.4%) | 0.53 |
| BMI (median, IQR) | 26.9 (23.9, 31.6) | 25 (22.7, 26.8) | 25.8 (23.3, 28.7) | 0.009 * |
| Chronic Liver Disease (N, %) | 9 (13.2%) | 5 (8.2%) | 14 (10.9%) | 0.41 |
| COPD (N, %) | 12 (17.7%) | 15 (24.6%) | 27 (20.9%) | 0.39 |
| Diabetes (N, %) | 17 (25.0%) | 16 (26.2%) | 33 (25.6%) | >0.95 |
| Heart Disease (N, %) | 36 (52.9%) | 32 (52.5%) | 68 (52.7%) | >0.95 |
| Kidney Disease (N, %) | 6 (8.8%) | 11 (18%) | 17 (13%) | 0.19 |
| ASA Class (N, %) | ||||
| 1–2 | 6 (13.6%) | 4 (14.3%) | 10 (13.9%) | >0.95 |
| 3+ | 38 (86.4%) | 24 (85.7%) | 62 (86.1%) | |
| Missing | 24 | 33 | 57 | |
| ECOG (N, %) | 0.02 * | |||
| 0–1 | 30 (81.8%) | 12 (52.2%) | 42 (70%) | |
| 2–4 | 7 (18.9%) | 11 (47.8%) | 18 (30%) | |
| Missing | 31 | 38 | 69 | |
| Cancer Type (N, %) | 0.04 * | |||
| Hematologic | 13 (19.1%) | 23 (37.7%) | 36 (27.9%) | |
| Solid | 39 (57.4%) | 30 (49.2%) | 69 (53.5%) | |
| Multiple | 16 (23.5%) | 8 (13.1%) | 24 (18.6%) | |
| Cancer Type (Specific) (N, %) | NA** | |||
| Breast | 9 (13.2%) | 3 (4.9%) | 12 (9.3%) | |
| Colorectal | 6 (8.8%) | 5 (8.2%) | 11 (8.5%) | |
| Gynecologic | 3 (4.4%) | 9 (14.8%) | 12 (9.3%) | |
| Hematologic | 13 (19.1%) | 23 (37.7%) | 36 (27.9%) | |
| Lung | 6 (8.8%) | 2 (3.3%) | 8 (6.2%) | |
| Other | 4 (5.9%) | 2 (3.3%) | 6 (4.7%) | |
| Upper Gastrointestinal | 6 (8.8%) | 4 (6.6%) | 10 (7.8%) | |
| Urologic | 3 (4.4%) | 4 (6.6%) | 7 (5.4%) | |
| Multiple | 16 (23.5%) | 8 (13.1%) | 24 (18.6%) | |
| Procedure (N, %) | 0.02 | |||
| Drainage | 10 (14.7%) | 22 (36.1%) | 32 (24.8%) | |
| Surgery | 22 (32.4%) | 13 (21.3%) | 35 (27.1%) | |
| No procedure | 36 (52.9%) | 26 (42.6%) | 62 (48.1%) |
Statistically significant p<0.05
Descriptive purposes only
In the univariable logistic regression model, ECOG score 2–4 (OR 0.26, 95% CI 0.08–0.81, P=0.02), hematologic malignancies (OR 0.44, 95% CI 0.19–1, P=0.049), gallbladder drainage (OR 0.27, 95% CI 0.10–0.74, P=0.011) were associated with a lower odd of RIOT ≤ 6 weeks (Table 2). In a multivariable logistic regression model evaluating RIOT ≤ 6 weeks, BMI, chronic kidney disease, ECOG, malignancy type, and type of intervention were included in the initial model based on P-value, however ECOG had >50% missing data and caused model instability. In the final model, patients with hematologic malignancies (compared to solid tumors) were 59% less likely to RIOT ≤ 6 weeks (OR 0.41, 95% CI 0.17–0.98, P=0.04) and patients who received drainage compared (compared to surgery) were 68% less likely (OR 0.32, 95% CI 0.13–0.81, P=0.02) (Table 3). In a univariable model predicting 90-day mortality, patients who had a drainage procedure had a 7.3-fold higher odds of death (OR 7.3, 95% CI 1.84,28.94, P=0.005) and patients who did not have any intervention had a 5.9-fold higher odds of death (OR 5.9, 95% CI 1.61, 21.4, P=0.007) compared to those who underwent surgery (Table 4).
Table 2.
Univariable Logistic Regression for RIOT <6-weeks
| Variable | Odds Ratio (OR) | 95% Confidence Interval (CI) | P-value |
|---|---|---|---|
| Age | 0.985 | 0.96, 1.01 | 0.27 |
| Sex (ref. female) | 1.071 | 0.53, 2.16 | 0.85 |
| Race (ref. White) | 0.726 | 0.32, 1.65 | 0.45 |
| BMI | 1.065 | 1.00, 1.13 | 0.051 |
| Chronic liver disease | 1.708 | 0.54, 5.41 | 0.36 |
| COPD | 0.657 | 0.28, 1.54 | 0.34 |
| Diabetes | 0.938 | 0.42, 2.07 | 0.87 |
| Heart disease | 1.020 | 0.51, 2.04 | >0.95 |
| Chronic kidney disease | 0.440 | 0.15, 1.27 | 0.13 |
| ASA classification III+ (ref. I-II) | 1.056 | 0.27, 4.13 | 0.94 |
| ECOG 2–4 (ref. 0–1) | 0.255 | 0.08, 0.81 | 0.02 |
| Hematologic malignancy | 0.435 | 0.19, 1.00 | 0.049 |
| Multiple malignancies | 1.538 | 0.58, 4.07 | 0.39 |
| Drainage procedure (ref. surgery) | 0.27 | 0.10, 0.74 | 0.011 |
| No procedure(ref. surgery) | 0.82 | 0.35, 1.92 | 0.064 |
Table 3.
Multivariable Logistic Regression for RIOT <6 weeks
| Variable | Odds Ratio (OR) | 95% Confidence Interval | P-value |
|---|---|---|---|
| BMI | 1.06 | 0.99, 1.13 | 0.12 |
| Chronic kidney disease | 0.41 | 0.13, 1.28 | 0.12 |
| Malignancy type (ref. solid) | |||
| Hematologic | 0.413 | 0.17, 0.98 | 0.04 |
| Multiple | 1.536 | 0.56, 4.24 | 0.41 |
| Procedure (ref. surgery) | |||
| Drainage | 0.32 | 0.13, 0.81 | 0.02 |
| No procedure | 0.88 | 0.36, 2.15 | 0.77 |
Table 4.
Univariable Logistic Regression for 90-day Mortality
| Variable | Odds Ratio (OR) | 95% Confidence Interval (CI) | P-value |
|---|---|---|---|
| Age | 1.01 | 0.98,1.04 | 0.48 |
| Male Sex (ref. female) | 0.74 | 0.34, 1.62 | 0.46 |
| Other Race (ref. White) | 1.82 | 0.77, 4.29 | 0.17 |
| BMI | 0.97 | 0.91, 1.04 | 0.41 |
| Chronic liver disease | 0.37 | 0.08, 1.72 | 0.20 |
| COPD | 1.26 | 0.51, 3.12 | 0.62 |
| Diabetes | 0.87 | 0.36, 2.09 | 0.75 |
| Heart disease | 1.0 | 0.47, 2.12 | >0.95 |
| Chronic kidney disease | 0.71 | 0.21, 2.32 | 0.57 |
| ASA classification III+ (ref. I-II) | 0.75 | 0.17, 3.25 | 0.70 |
| ECOG 2–4 (ref. 0–1) | 2.13 | 0.61, 7.39 | 0.24 |
| Hematologic malignancy | 0.76 | 0.31, 1.90 | 0.56 |
| Multiple malignancies | 1.14 | 0.42, 3.08 | 0.79 |
| Drainage procedure (ref. surgery) | 7.30 | 1.84, 28.94 | 0.005 |
| No procedure (ref. surgery) | 5.87 | 1.61, 21.37 | 0.007 |
Discussion
In this manuscript, we report that only half (52.7%) of oncology patients on active anti-cancer therapy who develop cholecystitis RIOT ≤ 6 weeks, and one quarter (24.8%) never return to treatment at all. In a multivariable model, patients with hematologic malignancies and those who had gallbladder drainage were 59% and 68% less likely to return to RIOT ≤ 6 weeks, respectively. A higher proportion of patients who underwent cholecystectomy returned to therapy within 6-weeks, compared to those who received gallbladder drainage (62.9% vs. 31.3%), however the majority of cholecystectomies were performed open (54.3%). Patients who underwent either drainage procedure or no procedural intervention were 7.3 and 5.9 times more likely to die within 90-days compared to those who underwent surgery. A low proportion of patients who received PCTs would go on to have definitive cholecystectomies (8.9%).
The relationship between hematologic malignancies and failure to RIOT is multifactorial and likely related to both cytopenias and immunosuppression and confounded by the heterogeneity of the patient population. While several studies have reported that patients with hematologic malignancies can safely undergo cholecystectomy with acceptable morbidity15,16, prior work from our group has shown that abdominal surgery in neutropenic patients carries a high morbidity (54.5%) and mortality (11.8%)11. In a retrospective case-control, single-institution series, Zhang et al showed a 16.7% mortality amongst hematopoietic stem cell transplant (HSCT) patients who develop acute cholecystitis17. Patients with moderate or severe acute cholecystitis had worse survival (94.1% vs. 77%, P=0.001) and non-relapse-related mortality (13.1% vs. 2.7%, P=0.011) measured up to 6.8 years after HSCT. We have a very active HSCT and cellular therapy program at our institution, which may underly some of the findings in our study, as a patient with indolent CLL will respond very different to cholecystitis than a neutropenic HSCT patient, but both are labeled as a hematologic malignancy using claims data. There is likely an element of selection bias wherein sicker patients, less suitable for surgery, underwent gallbladder drainage and were not subjected to a potentially morbid and/or futile operation. Clearly, surgery was appropriately avoided among the cohort of 32 patients who would never return to therapy due to either death in hospital, transition to hospice/DNR, or inability to tolerate additional chemotherapy.
This study also provides important information about the utility of PCT in cancer patients, as this was the majority of drainage procedures. Anecdotally, providers may use PCT as a temporizing measure with the intention to provide source control and expedite return to treatment. However, of the 32 patients who had a PCT placed, less than 15% returned to their intended treatment within 6-weeks. This may be due to features of the underlying malignancy and patient fitness, as patients who did not RIOT ≤ 6 weeks had ECOG scores. This is corroborated by the recent multi-institutional study by Johnson et al12, which found that before propensity score matching, patients with cancer and acute biliary disease were older, had more comorbidities, and a higher severity of gallbladder disease12. However, we must also consider the complications and second-order effects of PCT as a procedure. In the multicenter randomized CHOCOLATE trial, in which high-risk patients were randomized to undergo PCT or laparoscopic cholecystectomy, investigators found a 53% higher rate of major complications, including reinterventions and recurrent biliary disease in the PCT arm18. Most recently, in a retrospective cohort study using the TriNetX database, Podder et al19 found that 41% of patients who receive a PCT require at least one additional biliary intervention within 1-year. Understandably, reinterventions, recurrent biliary symptoms, and complications from the procedure itself are factors that may prevent patients from returning to their oncologic therapy.
Similarly, our data do not support PCT as a bridge to interval cholecystectomy as only 8.9% would go on to have definitive cholecystectomy. In the same retrospective study by Podder et al19, only 43% of patients who received a PCT would undergo interval cholecystectomy and rates are as low as 28% in other publications20–22. Our low rate of interval cholecystectomy is likely confounded by the severity of patient illness, as demonstrated by the significantly higher odds of 90-day mortality for both the cohorts receiving gallbladder drainage (OR 7.3) and no procedural intervention (OR 5.9). We must be honest with patients that PCT is a destination treatment for most patients based on multiple studies. There may also be an association with survival by pursuing surgery over non-operative management. In a recent comparative effectiveness study by Acker et al22 analyzing Medicare beneficiaries with multiple comorbidities treated for cholecystitis either operatively or non-operatively (which included PCT placement), risk-adjustment using inverse propensity weights demonstrated that operative treatment was associated with significantly lower 30-day mortality, 30- and 90-day readmissions, and 30- and 90-day ED re-visit rates compared to non-operative treatment 22. Unfortunately, definitive recommendations cannot be drawn from retrospective series and further study is needed.
In terms of limitations, using ICD-9/10 codes for cohort creation mis-identified a large number of patients (Figure 1). This is consistent with prior literature describing a κ between chart and administrative database coding for certain conditions as low as 0.1323. 24,25This study is a retrospective series from a standalone cancer center, which limits its generalizability to other centers and is inherently influenced by selection bias in terms of treatment allocation. Our institution serves a very niche cohort of patients, and this study captures patients who remain local during systemic therapy and have to the ability to return to the UCC for complications. As such, we are likely not capturing all patients. However, this is one of the largest granular series on cholecystitis in cancer patients and provides important information about the impact of cholecystitis on cancer-directed treatment and the realistic outcomes of various management strategies. Our low rate of conversion of PCT to interval cholecystectomy and rate of open cholecystectomy is surprising and may represent the underlying fitness of the patients, local institutional practices, or surgeon judgement. EGS call is taken by a pool of general surgery-trained surgical oncologists and most of the cholecystectomies were performed by hepatopancreatobiliary surgeons. In terms of the low rate of conversion of PCT to cholecystectomy, many of these patients died in the hospital, were transitioned to hospice/made DNR, or could not tolerate further chemotherapy. As a tertiary medical center, it is possible some patients had further management of their cholecystitis at outside facilities after PCT placement, which were not captured in our data and may also partially explain our low rate of interval cholecystectomy. We dichotomized our end-points (RIOT 6-weeks and 90-day mortality) as time-to-event analyses in this cohort was not interpretable. This type of analysis would require a competing risk model (death without RIOT as a competing event) with a time varying covariate for procedure because at the time of EGS diagnosis, the timing and eventual treatment is unknown. Coupled with this being an emergency event, a shorter-term endpoint was deemed more appropriate. In terms of the modeling for 90-day mortality, as there were only 17 events, running a multivariable-adjusted model would have led to overfitting. Moreover, the only other variable with a P-value of <0.2 was non-white race.
Conclusion:
Only one half of patients who develop cholecystitis while on treatment for cancer RIOT ≤ 6-weeks, and one-quarter will never return to treatment. Patients with hematological malignancies and those who have PCTs placed are less likely to return to RIOT ≤ 6-weeks. Patients who undergo cholecystectomy are more likely to RIOT ≤ 6-weeks, but this is likely multifactorial as patients who undergo either gallbladder drainage or no procedural intervention have a higher odd of 90-day mortality. There is a low rate of interval cholecystectomy after PCT and PCT often serves as destination therapy and does not bridge patients back to cancer-directed treatment.
Supplementary Material
Funding/Support:
This work was supported in part by the Marie-Josée and Henry R. Kravis Center for Molecular Oncology, the National Cancer Institute (NCI) Cancer Center Support Grant [P30CA008748]. Funding sources had no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.
Footnotes
Conflicts of Interest (COI)/Disclosure: The authors have no relevant financial disclosures.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Hester CA, Karbhari N, Rich NE, et al. Effect of fragmentation of cancer care on treatment use and survival in hepatocellular carcinoma. Cancer. 2019;125(19):3428–3436. doi: 10.1002/CNCR.32336 [DOI] [PubMed] [Google Scholar]
- 2.Choi DW, Kim SJ, Kim DJ, Chang YJ, Kim DW, Han KT. Does fragmented cancer care affect survival? Analysis of gastric cancer patients using national insurance claim data. BMC Health Serv Res. 2022;22(1):1566. doi: 10.1186/S12913-022-08988-Y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Jolissaint JS, Lobaugh SM, Goldman DA, et al. Frequency and Natural History of Emergency General Surgery Conditions in Cancer Patients: A SEER-Medicare Population Analysis. Ann Surg. 2026;5(283):783–790. https://journals.lww.com/annalsofsurgery/fulltext/9900/frequency_and_natural_history_of_emergency_general.1092.aspx [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Nelson DB, Rice DC, Mitchell KG, et al. Return to intended oncologic treatment after surgery for malignant pleural mesothelioma. Journal of Thoracic and Cardiovascular Surgery. 2019;158(3):924–929. doi: 10.1016/j.jtcvs.2019.02.129 [DOI] [PubMed] [Google Scholar]
- 5.Mickel TA, Kutlu OC, Silberfein EJ, et al. Factors associated with inability to return to intended oncologic treatment in pancreatic cancer. Am J Surg. 2022;224(1):635–640. doi: 10.1016/j.amjsurg.2022.02.058 [DOI] [PubMed] [Google Scholar]
- 6.Kim Y, Go TH, Jang J, et al. Survival impact of adherence to tyrosine kinase inhibitor in chronic myeloid leukemia. Korean J Intern Med. 2021;36(6):1450–1458. doi: 10.3904/KJIM.2021.158 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Długosz-Danecka M, Szmit S, Ogórka T, Skotnicki AB, Jurczak W. The average relative dose intensity of R-CHOP is an independent factor determining favorable overall survival in diffuse large B-cell lymphoma patients. Cancer Med. 2019;8(3):1103. doi: 10.1002/CAM4.2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Scott JW, Olufajo OA, Brat GA, et al. Use of national burden to define operative emergency general surgery. JAMA Surg. 2016;151(6):1–7. doi: 10.1001/jamasurg.2016.0480 [DOI] [PubMed] [Google Scholar]
- 9.von Kroge PH, Duprée A, Mann O, et al. Abdominal emergency surgery in patients with hematological malignancies: a retrospective single-center analysis. World Journal of Emergency Surgery 2023 18:1. 2023;18(1):12-. doi: 10.1186/S13017-023-00481-Z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Mallampalli GM, Campbell DB, Srinivas S, et al. Oncology and Acute Care Surgery: Examining Outcomes in Cancer Patients Requiring Emergent Laparotomy. Journal of Surgical Research. 2025;310:80–88. doi: 10.1016/j.jss.2025.03.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jolissaint JS, Harary M, Saadat LV., et al. Timing and Outcomes of Abdominal Surgery in Neutropenic Patients. Journal of Gastrointestinal Surgery. 2019;23(4):643–650. doi: 10.1007/s11605-018-04081-0 [DOI] [PubMed] [Google Scholar]
- 12.Johnson PL, Williams JE, Schaefer SL, et al. Variation in acute cholecystitis outcomes and processes of care in patients with cancer. Surgery (United States). 2025;179. doi: 10.1016/j.surg.2024.07.078 [DOI] [PubMed] [Google Scholar]
- 13.Guerra-Londono CE, Schreck A, Muthukumar A, Guerra-Londono JJ. Return to intended oncologic treatment: Definitions, perioperative prognostic factors, and interventions. Best Pract Res Clin Anaesthesiol. 2025;39(1):14–22. doi: 10.1016/J.BPA.2025.03.009 [DOI] [Google Scholar]
- 14.Aloia TA, Zimmitti G, Conrad C, Gottumukalla V, Kopetz S, Vauthey JN. Return to intended oncologic treatment (RIOT): a novel metric for evaluating the quality of oncosurgical therapy for malignancy. J Surg Oncol. 2014;110(2):107–114. doi: 10.1002/jso.23626 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Stewart CL, Tran TB, Nguyen A, et al. Cholecystectomy in patients with hematologic malignancies. Am J Surg. 2022;223(6):1157–1161. doi: 10.1016/j.amjsurg.2021.10.037 [DOI] [PubMed] [Google Scholar]
- 16.George J, Cohen M, Whitney S, Fennern E, Divino C. Biliary Disease in Immunocompromised Patients: a Single-Center Retrospective Analysis. Journal of Gastrointestinal Surgery. 2019;23(8):1589–1592. doi: 10.1007/s11605-018-04077-w [DOI] [PubMed] [Google Scholar]
- 17.Zhang Z, He Y, Zhu XL, et al. Acute Cholecystitis Following Allogeneic Hematopoietic Stem Cell Transplantation: Clinical Features, Outcomes, Risk Factors, and Prediction Model. Transplant Cell Ther. 2021;27(3):253.e1–253.e9. doi: 10.1016/j.jtct.2020.12.016 [DOI] [PubMed] [Google Scholar]
- 18.Loozen CS, Van Santvoort HC, Van Duijvendijk P, et al. Laparoscopic cholecystectomy versus percutaneous catheter drainage for acute cholecystitis in high risk patients (CHOCOLATE): multicentre randomised clinical trial. BMJ. 2018;363:3965. doi: 10.1136/BMJ.K3965 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Podder S, Lung K, Ibrahim G, et al. Barriers to interval cholecystectomy following percutaneous cholecystostomy in patients with acute calculous cholecystitis. Surg Endosc. 2025;39(12). doi: 10.1007/S00464-025-12161-X [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sadaka AH, Tseng JF, Itani KMF. Indications for and Optimal Management of Percutaneous Cholecystectomy Drainage. JAMA Surg. Published online September 3, 2025. doi: 10.1001/jamasurg.2025.3260 [DOI] [PubMed] [Google Scholar]
- 21.Colonna AL, Griffiths TM, Robison DC, et al. Cholecystostomy: Are we using it correctly? Am J Surg. 2019;217(6):1010–1015. doi: 10.1016/j.amjsurg.2019.04.002 [DOI] [PubMed] [Google Scholar]
- 22.Acker RC, Ginzberg SP, Sharpe J, et al. Operative vs Nonoperative Treatment of Acute Cholecystitis in Older Adults with Multimorbidity. JAMA Surg. 2025;160(6):656–664. doi: 10.1001/jamasurg.2025.0729 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Quan H, Li B, Duncan Saunders L, et al. Assessing Validity of ICD-9-CM and ICD-10 Administrative Data in Recording Clinical Conditions in a Unique Dually Coded Database. Health Serv Res. 2008;43(4):1424. doi: 10.1111/J.1475-6773.2007.00822.X [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Lyu HG, Stein LA, Saadat LV., Phicil SN, Haider A, Raut CP. Assessment of the Accuracy of Disease Coding Among Patients Diagnosed With Sarcoma. JAMA Oncol. 2018;4(9):1293. doi: 10.1001/JAMAONCOL.2018.2979 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Sweeney SM, Hamadeh HK, Abrams N, et al. Challenges to Using Big Data in Cancer. Cancer Res. 2023;83(8):1175. doi: 10.1158/0008-5472.CAN-22-1274 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.