Impact of Early Palliative Care on end of life in Patients with Advanced Biliary Tract Cancer

Keywords

Cholangiocarcinoma • Biliary tract cancer • Gallbladder carcinoma • Palliative care • Survival

Abbreviations

BTC: Biliary Tract Cancer; aBTC: advanced Biliary Tract Cancer; CCA: Cholangiocarcinoma; iCCA: intra Hepatic Cholangiocarcinoma; eCCA: extra Hepatic Cholangiocarcinoma; pCCA: peri Hilar Cholangiocarcinoma; ED: Emergency Department; ICU: Intensive Care Unit; OS: Overall Survival; PCU: Palliative Care Unit; PCT: Palliative Care Teams; PCR: Palliative Care Referral; PS: Performans Status

INTRODUCTION

Biliary Tract Cancers (BTC) is rare primary liver cancers but its incidence is increasing. Prognosis remains poor with Overall Survival (OS) reaching 6 to 11 months [1,2]. That poor prognosis is attributed to challenges in early detection (due to delayed symptoms), low opportunity for radical resection and limited response to chemotherapy and/or radiotherapy. Indeed, only 20% to 30% of patients diagnosed with BTC have access to surgery, which is the only curative-intent treatment [3]. Therefore, most of patients receive palliative systemic therapy (chemotherapy, targeted therapy, and immunotherapy) or exclusive supportive care.

Palliative Care Teams (PCT) is composed of several actors making possible a multidisciplinary and global comprehensive care: physical, emotional, social (family caregivers…), spiritual (meaning, dignity, faith…), informational (prognosis, illness understanding…) [4]. Palliative care (PC) and systemic palliative therapy are complementary approaches that should be associated early for more benefits. In 2017, the American Society of Clinical Oncology (ASCO) published recommendations about palliative care, specifying that palliative care should be introduce early in cancer management, at diagnosis and with the beginning of systemic therapy [5].

Despite these recommendations, and probably due to a lack of palliative care teams, patients’ referral to PCT in daily practice is often late (eg: no more therapeutic options, general condition too weak).

Recent studies demonstrated the positive impact of PC on aggressiveness of end-of-life care that includes chemotherapy, intensive care unit admission, Emergency Department (ED) visit and acute unit hospitalizat6ion, during last month of life [6]. In pancreatic cancer in a retrospective study the positive impact of a PC consultation on these above mentioned end-of-life indicators [7]. These results were confirmed in a prospective multicentric randomized controlled Italian trial including 206 patients with metastatic pancreatic cancer [8].

Recently, the concept of PC has evolved with the issue of the timing of palliative care referral (PCR), and the benefits of an early palliative care. An Australian retrospective study highlighted the use of early PC to improve aggressiveness of end-of-life care, compared to late PC (>90 days before death) in patients with pancreatic cancer [9].

Moreover, a retrospective study among patients in advanced lung cancer shown that early PC (30 to 365 after diagnosis) was significantly associated with an increase of OS (aHR, 0.47; 95% CI, 0.45-0.49). In 2021, a randomized study of advanced esophagogastric cancer also demonstrated an OS improvement in case of early interdisciplinary supportive care.

In the context of aBTC, which prognosis is poor with frequent uncomfortable symptoms as jaundice, anorexia, asthenia, abdominal pain, weight loss, psychological suffering; palliative care team have a crucial role to provide a multidisciplinary take of care. To our knowledge, no study assessed early PCR in aBTC patients. We conducted a retrospective multicentric study to assess the relationship between early PCR and survival (primary outcome) and to explore the impact of early PCR on treatment aggressiveness near the end of life [10-12].

Materials and Methods

We conducted a retrospective multicentric study including patients with a diagnosis of aBTC (cholangiocarcinoma and gallbladder) between January 1, 2013 and December 31, 2019. We included patients of 6 University hospitals of Eastern France (Nancy University Hospital, Besançon University Hospital, Dijon University Hospital, Reims University Hospital, Strasbourg University Hospital, and Strasbourg ICANS Centre for the Treatment of Cancer). Follow-up data were available until January 2021. Detailed clinical and pathology data were collected for all patients through a custom-made questionnaire providing epidemiological informations, data about aBTC (age at diagnosis, risk factors, treatment) and palliative care support (time to BSC referral, end-of-life acute events and condition). When necessary, we consulted medical files on site. Ethics approval was obtained from the Ethics Committee of Nancy CHU September 8, 2020, Licence number 272. Patients received an information letter and had 30 days to ask their withdrawal from the study. After this delay, they were considered as participants.

Cohort selection

Patients were eligible if they were ≥ 18 years old, had and histologically confirmed, locally advanced or metastatic cholangiocarcinoma or gallbladder adenocarcinoma diagnosed between January 1, 2013 and December 31, 2019. Patients were included regardless their Eastern Cooperative Oncology Group (ECOG) performance status

Patients were excluded if they died within a week after diagnosis and if insufficient clinical data were available.

Exposure

PC was defined as a specialist-delivered palliative care encounter received inpatient or outpatient setting after aBTC diagnosis. It could be with a mobile PCT, in hospitalization (palliative care units, PCU) or in consultation. Data were collected in electronic medical records (DxCare software).

Study objectives

Primary outcome was OS after aBTC diagnosis according to PCR. We secondary evaluated the impact of early PCR on the aggressiveness of end-of-life care, which included place of death (at home with home hospitalization, at home without home hospitalization, intensive care unit, conventional acute unit, palliative care unit), last-month chemotherapy, last-month ED visit and last-month intensive care unit hospitalization. Continuous variables were summarized by descriptive statistics (number of cases, mean, median, and quartile) and categorical variables were summarized with percentages. We also compared the most frequently reported reasons to referral and proposals made by PCT.

Statistical analysis

Characteristics of sample were described by percentage for categorical variables and median with inter-quartile range (Q1-Q3) for continued variables. OS was calculated from the date of diagnosis of advanced BTC until the date of death or date of last follow-up and Kaplan-Meier curve was realized for each group. A second analysis was made by calculating the time from the date of first meeting with PC until the date of death or date of last follow-up. To address potential selection bias, we have planned to make a propensity score of receiving palliative care, composed of metastatic stage, age (>75 years old), performans status OMS, Charlson score (>3). This approach minimized selection bias by comparing patients with a similar likelihood of receiving PC. Due to heterogeneity between study groups, propensity score has not been achieved. Clinical characteristics among these four groups were too different to produce a reliable comparison. The significant threshold was fixed at 5%. All analyses were performed with SAS® software version 9.4 (SAS Institute, Cary, NC, USA).

Results

Patients’ characteristics

Two hundred patients with aBTC were included in our study. Median age at aBTC diagnosis was 67.6 years old (20-91 years old). 61% of the patients were men. Most of the tumours were intrahepatic cholangiocarcinoma (iCCA) (51%), then perihilar (pCCA) (24%), gallbladder (18%) and distal CCA (7%). 62.5% patients had distant metastasis at diagnosis and 37.5% were locally advanced. About well-known BTC risk factors, 27% of patients were smokers and 18% had a heavy alcohol consumption, 19% were obese (body mass index ≥ 30 kg/m2), 31% had type II diabetes, 14% had a cirrhosis. Only 2 patients had a primary sclerosing cholangitis. The most common symptoms at diagnosis were abdominal pain (40%), jaundice (33%), asthenia (31%) and anorexia (28%). A quarter (26%) had a prior curative-intent treatment before their relapse (surgery or locoregional treatment) and then a non-curative recurrence. Systemic palliative chemotherapy was prescribed for 82.5% of the patients, and 17.5% received exclusive supportive care (Table 1).

Among the 165 patients who received chemotherapy, the most frequent (67%) first-line regimen was Gemcitabine+Oxaliplatine (GEMOX). Other protocols administred as first line palliative therapy where Gemcitabine+Cisplatin (14%) and gemcitabine alone (10%). More than half of the patients received a second-line chemotherapy (56%), mainly 5FUbased regimen (54% received FOLFIRI).

Palliative care

Most of included patients (56%) met the PCT at least once (Table 1). The PCT were constituted of at least one specialized doctor and a nurse. The mean time to first PCR was more than 6 months after diagnosis (9.7 months) (Table 2).

In most of the 6 centres, patients benefit of supportive care through an encounter with mobile PCT (65%), less through a hospitalization in PC units (26%) or a consultation (9%). In one of the centres (ICANS), we notice there was hospital day care that allows a global take of care, with participation of several stakeholders during one day.

This support by PCT mainly often results of a request by the oncologist. The most frequent reasons for referral were asthenia (58%), decision for place of care (41%) and pain (40%). PCR interventions were “decision for becoming and place of care” (54%), opioids treatment (49%), anxiolytics (36%), mouth care (36%) (Not mutually exclusive).

Impact of palliative care on overall survival

Among the 200 included patients, 188 (94%) died at the end of data collection. 87 (43.5%) did not received PC, 30 (15%) had early PC (<3 months after diagnosis), 20 (10%) had PC 3 to 6 months after diagnosis, and 63 (31.5%) had late PC (> 6 months). Median OS were respectively in the four groups 12.4 months (no PC), 3.0 months (PC < 3 months), 6.4 months (PC 3-6 months) and 16.0 months (PC >6 months) (Table 3).

Figure 1 represents OS with or without PC. OS was not significantly different between PC and no PC groups (p=0.34). Figure 2 represents survival according to the first PCR. No significant difference was observed between groups (p=0.79). Time between PCT contact and death were 0.9 months (PC < 3months), 1 month (PC 3-6 months) and 1.3 months (PC >6 months) (Figures 1 and 2).

Figure 1. Association between survival and palliative care: Palliative care vs. no palliative care.

Figure 2. Survival after first meeting with palliative care team.

Impact of palliative care on aggressiveness in end-of-life care

Place of Death: Place of death seemed to be influenced by PC (not significant), especially when PCR was early (<3 months after diagnosis). PC was associated with less death in conventional acute unit, intensive care unit and ED (respectively 72% no PC vs. 21% early PC and 8% no PC vs. 0% early PC) and more death in palliative care unit (Table 3).

Last-month Chemotherapy: PC was associated with less chemotherapy during last month of life (non-significant): 36% (no PC), 30% (PC <3months), 10% (PC 3-6months), 10% (PC >6months).

Last-Month Emergency Department Visit: Results were similar for ED visits near death. PC was associated with less visits in ED, However the support of early PCR was not significant (20% PC<3months, 15% PC 3-6months, 7% PC >6months; versus 35% no PC).

Last-Month Intensive Care Unit Hospitalization: Hospitalizations in intensive care unit seemed to be impacted by PC: 10 patients in the no PC group were admitted in ICU versus 1 patient in the PC group.

Discussion

Because of its advanced/metastatic stage at diagnosis and its poor prognosis, the management of aBTC should be inseparable from the initiation of early palliative care, as soon as systemic palliative therapy is initiated.

To our knowledge, the present study is the first which provides real-world data about PC settings and which assess the impact of this support on survival in patients with aBTC.

Indeed, there is scarce literature available about biliary tract cancer and PC. Therefore, we will compare our data with available literature regarding advanced/metastatic pancreatic ductal adenocarcinoma, which prognosis is similar.

In our study, 17% of patients were not able to received systemic treatment, which is different from other published studies which included 3650 iCCA patients). Such a difference might be partly explained, by the fact that in this above mentioned study, all patients with advanced or metastatic BTC were included, even those without accessible histological evidence, making it feared that their general condition was too poor at diagnosis to obtain histological evidence. In the same way, more than a half (56%) of chemotherapy-treated patients had a second-line chemotherapy, which is more than other studies (less than one third) [13,14].

Our main objective was to assess if early PC could improve OS. We found no evidence that OS was better in PC group in comparison to no PC group. It could be explained by the heterogeneity between groups. Indeed, groups were not comparable: patients who were referred earlier to PCT had more comorbidities and more advanced BTC. The group of patients who met PCT early after aBTC diagnosis had the worst survival due to several poor prognosis factors: high performans status (≥ 2 for 62% of them), more comorbidities (Charlson Index Score ≥2 for 47% of them), and only one third of them was able to receive palliative chemotherapy. Consequently, causes of death seem to be also different between groups. Because patients from early PC groups had more frailties and comorbidities, related-cause of death were mainly due to deterioration of general conditions (93% for early PC vs. 59% no PC). Differences in the rate of acute events were also noticed, as sepsis (7% early PC vs. 24% no PC) or gastro-intestinal bleeding (0% early PC vs. 4% no PC). On the contrary, patients who never receive PC died more from acute complication as sepsis or bleeding, with more hospitalizations in intensive care unit.

Differences are also due to the retrospective character of our study, which reflects our daily clinical practice, with patients referred to PCT 0.9 to 1.3 months before death, usually when systemic treatments are no more possible. The early PC group is representative of these practices: patients who were referred early to PCT after aBTC diagnosis had symptomatic advanced or metastatic BTC cancer and presented altered general conditions. This is probably the main explanation for poor OS (3 months) in that group. In such situations, PCT cannot build strong relationship with patient and their family. However, important times for exchanges should be established, as interruption of systemic chemotherapy, end-of-life preferences, and meeting family members to have a global action. These results are similar to a French study published in 2017, about all patients addressed to PCT of 25 Parisians hospitals: Median survival time after first PCT meeting was 31 days, and 18% of the patients first met PCT 3 days before death [15].

In 2017, the American Society of Clinical Oncology (ASCO), published recommendations about palliative care, and especially mentioned that patients with advanced cancer should receive dedicated palliative care services early in the disease course and concurrent with active treatment. In our study, 56% of the patients have met at less once the PCT, which is similar that a retrospective study in pancreatic cancer patients (52%). However, time-to-referral seems too late for impacting prognosis. Average time between diagnosis and PCR was 9.7 months, although median OS fluctuates from 3 to 16 months. Among 11% of the patients who met PCT, this meeting happened during the last hospitalization (a few days before death), which is too late.

The most common reasons for PCR in our study were asthenia (58%), “decision for becoming and place of care” (41%), pain (40%), psychological symptoms (12%) and then following the cessation of active therapeutics (11%). As a comparison demonstrated in their study that the 3 major reasons for PCR (for all cancer types) were pain (57%), early encounter (24%), decision for place of care (18%) and then in 6th position “decision to withhold or withdraw treatments” (9.3%). In our study, patients were mostly referred to PCT for difficulties for end-of-life management. Late PCR does not allow PCT to be a part of the decision about active treatment, which should be discussed in a multidisciplinary team.

In our study, meeting PCT is associated with an impact on place of death (Table 3). When they received early PC, fewer patients died in conventional, emergency or intensive departments. In a retrospective study published in 2015, conducted in more than 500 000 patients with advanced cancer, proportion of death in PC units was 10% showed that this proportion reached 30% when patients were referred to PCT. It reflects the positive impact of PC on the place of death, with fewer deaths in acute care units. Death in appropriate and dedicated department is important for a better management of end-of-life. In palliative care units or home hospitalization, PCT are familiar with those situations, and are able to anticipate future needs for patients, and to discuss of their preferences in order to head off difficult situations of end-of-life. In a retrospective study performed in patients with pancreatic adenocarcinoma, chemotherapy was prescribed during last month in 17% in case of early PCR and 25% in case of late PCR (not significant, p=0,22). In our study, 18% of patients received palliative chemotherapy within the last month of life (36% no PC, 10%-30% with PC), which is similar to previous studies. PC seems to reduce end-life chemotherapy. Chemotherapy at end-of-life does not bring any benefits for patients, leading to a worsening of the side effects of systemic drugs (eg: asthenia, nausea, infections leading to hospitalizations). Such situations could be avoided with early PCT meeting [16].

In the Australian study focusing on advanced/metastatic pancreatic cancer patients early PC allows a significant reduction of last-month visit in ED (24%) in comparison to late PSC (42%) (p=0,003). In our study, in the PC group, ED visit proportions varied from 7% to 20%, compared with 35% in the no PC group. Our data also suggests that PC reduces ED visits during the last month of life.

In 2005, Earl et al. proposed some objectives to evaluate aggressiveness in end-of-life care

1. Less than 10% of patients should receive chemotherapy in the last 14 days of life. In our study we collected last-month chemotherapy rate (36% without PC vs. 10% to 30% with PC)

2. Less than 4% of emergency room visits. In our study we had 35% no PC group and 7% to 20% PC group

3. Less than 17% death in an acute care hospital. We had 8% death in ICU in non-PC group vs. 0% to 5% in PC group. In conventional acute unit we had 73% in non PC group vs. 21% to 25% in PC group.

4. Less than 4% of admission in ICU. In our study this objective is achieved for the PC group (0%-2%) but no for the no PC group (13%).

One major limitation of our study is its retrospective character leading to heterogeneity into groups and therefore a bias especially for overall survival in the four groups.

Another important limitation of our study is the differences of accessibility and support by the PCT. Indeed, although recommendations are established, referral to PCT is not standardized in France. Human and financial resources allocated to palliative care are different in France and vary from one center to another. This is well illustrated in our work by the presence of a day hospital dedicated to palliative care in only one of the 6 centers studied (Strasbourg, ICANS).

This study provides important data about end-of-life management, and partnerships with PCT for patients suffering from aBTC. Although no evidence for improvement of OS with early PCR was demonstrated, PC is introduced too late in the patient’s care pathways to improve prognosis. However, PC have a positive impact on end-of-life with limitation of futile systemic chemotherapy, fewer visit in ED, fewer hospitalization in ICU, and fewer death in a conventional acute unit or ICU.

Conclusion

This study provides important data about end-of-life management, and partnerships with PCT for patients suffering from aBTC. Although no evidence for improvement of OS with early PCR was demonstrated, PC is introduced too late in the patient’s care pathways to improve prognosis. However, PC have a positive impact on end-of-life with limitation of futile systemic chemotherapy, fewer visit in ED, fewer hospitalization in ICU, and fewer death in a conventional acute unit or ICU.

Despite ASCO recommendations of 2017 in favour of early PCR in cancer course (concomitant to active therapy), we observed that patients are mainly referred to PCT at end-of-life clinical in daily practise. Efforts should be done for a better education and access about PCR role in a multidisciplinary approach in cancer. Introduction of early PC should endorse a central role in aBTC cancer, with many positive consequences on end-of-life preferences, discontinuation of useless chemotherapy, better management of symptoms and family needs. PC should also evolve for each patient according to disease, prognosis, comorbidities, and symptoms, moving towards a “personalized early PC”.

Declarations

Ethics approval and consent to participate

Ethics approval was obtained from the Ethics Committee of Nancy CHU (September 8, 2020), Licence number 272.

Consent for publication

Not applicable for patients.

Availability of data and material

The data that support the findings of this study are available on request from the corresponding author (AL). The data are not publicly available due to information that could compromise research participant privacy.

Code availability: All analyses were performed with SAS® software version 9.4 (SAS Institute, Cary, NC, USA).

Conflicts of interest

The authors declare that they have no competing interests (files attached)

Funding

This research was supported by CHRU of Nancy. No funding

Author's contributions

Supervision, concept and design: A. Lopez. Data acquisition: M. Miralles, C. Borg, S. Manfredi, A. Minello, O. Bouché, M. Tambou, D. Mutter, P. Chiappa, J-E. Kurtz, A. Lopez. Statistical analysis: A. Luc, C. Baumann. Manuscript writing and editing: M. Miralles, M. Muller, C Borg, S. Manfredi, A. Minello, O. Bouché, M. Tambou, D. Mutter, P. Chiappa, J-E. Kurtz, A Luc, C. Baumann, A. Lopez. Drafting of the manuscript: A. Lopez, M. Miralles, and M Muller. Final approval of manuscript for all authors.

Acknowledgement

The authors would like to acknowledge the statistician team of University Hospitals of Nancy, especially Amandine Luc and Dr Baumann.

Thanks to the 6 hospitals of Eastern France that agreed to participate to this study.

Consent to participate

Permission to publish the case anonymously was obtained from patients.

References

1. Al Mahjoub, Aimen, Bouvier Véronique, Menahem Benjamin and Bazille Céline, et al. “Epidemiology of Intrahepatic, Perihilar, and Distal Cholangiocarcinoma in the French Population.” Eur J Gastroenterol Hepatol 31(2019): 678‑684. 

   [Crossref] [Google Scholar] [PubMed]

2. Flemming, Jennifer, Jina Zhang-Salomons, Sulaiman Nanji and Christopher M Booth. “Increased Incidence but Improved Median Overall Survival for Biliary Tract Cancers Diagnosed in Ontario from 1994 Through 2012: A Population-Based Study.” Cancer 122(2016): 2534‑2543.

   [Crossref] [Google Scholar] [PubMed]

3. Rizvi, Sumera, Shahid A Khan, Christopher Hallemeier and Robin Kelley, et al. “Cholangiocarcinoma-Evolving Concepts and Therapeutic Strategies.” Nat Rev Clin Oncol 15(2018): 95‑111.

   [Crossref] [Google Scholar] [PubMed]

4. Hui, David, Breffni L Hannon, Camilla Zimmermann and Eduardo Bruera. “Improving Patient and Caregiver Outcomes in Oncology: Team-Based, Timely, and Targeted Palliative Care.” CA Cancer J Clin 68(2018): 356‑76. 

   [Crossref] [Google Scholar] [PubMed]

5. Ferrell, Betty, Jennifer S Temel, Sarah Temin and Thomas J Smith. “Integration of Palliative Care into Standard Oncology Care: ASCO Clinical Practice Guideline Update Summary.” J Oncol Pract 13(2017): 119‑21.

   [Crossref] [Google Scholar] [PubMed]

6. Earle, Craig, Bridget A Neville, Mary Beth Landrum and Jeffrey M Souza, et al. “Evaluating Claims-Based Indicators of the Intensity of End-of-Life Cancer Care.” Int J Qual Health Care 17(2005): 505‑509.

   [Crossref] [Google Scholar] [PubMed]

7. Jang, Raymond, Monika K Krzyzanowska, Camilla Zimmermann and Nathan Taback, et al. “Palliative Care and the Aggressiveness of End-of-Life Care in Patients with Advanced Pancreatic Cancer.” J Natl Cancer Inst 107(2015): 424.

   [Crossref] [Google Scholar] [PubMed]

8. Maltoni, Marco, Emanuela Scarpi, Monia Dall'Agata and Vittorina Zagonel, et al. “Systematic Versus On-Demand Early Palliative Care: Results from a Multicentre, Randomised Clinical Trial.” Eur J Cancer 65(2016): 61‑68.

   [Crossref] [Google Scholar] [PubMed]

9. Michael, Natasha, Greta Beale, Clare O’Callaghan and Adelaide Melia, et al. “Timing of Palliative Care Referral and Aggressive Cancer Care toward the End-Of-Life in Pancreatic Cancer: A Retrospective, Single-Center Observational Study.” BMC Palliat Care 18(2019): 13.

   [Crossref] [Google Scholar] [PubMed]

10. Sullivan, Donald, Benjamin Chan, Jodi A Lapidus and Linda Ganzini, et al. “Association of Early Palliative Care use with Survival and Place of Death among Patients with Advanced Lung Cancer Receiving Care in the Veterans Health Administration.” JAMA Oncol 5(2019): 1702-1709.

    [Crossref] [Google Scholar] [PubMed]

11. Lu, Zhihao, Yu Fang, Chang Liu and Xiaotian Zhang, et al. “Early Interdisciplinary Supportive Care in Patients with Previously Untreated Metastatic Esophagogastric Cancer: A Phase III Randomized Controlled Trial.” J Clin Oncol 39(2021): 748‑756.

    [Crossref] [Google Scholar] [PubMed]

12. Blechacz, Boris and Gregory J Gores. Cholangiocarcinoma: “Advances in Pathogenesis, Diagnosis, and Treatment.” Hepatology 48(2008): 308‑321.

    [Crossref] [Google Scholar] [PubMed]

13. Neuzillet, Cindy, Corinne Emery, Clément Teissier and Stéphane Bouée et al. “Epidémiologie et parcours de soins des cholangiocarcinomes intrahépatiques (iCCA) en France: données de vie réelle issues du PMSI.” Revue d’Épidémiologie et de Santé Publique 68(2020): S77.

    [Crossref] [Google Scholar]  

14. Lamarca, Angela, Richard A Hubner, David Ryder and Juan W Valle. “Second-line chemotherapy in advanced biliary cancer: a systematic review.” Ann Oncol 25(2014): 2328‑2338.

    [Crossref] [Google Scholar] [PubMed]

15. Vinant, Pascale, Ingrid Joffin, Laure Serresse and Sophie Grabar, et al. “Integration and Activity of Hospital-Based Palliative Care Consultation Teams: The Insight Multicentric Cohort Study.” BMC Palliat Care 16(2017): 36.

    [Crossref] [Google Scholar] [PubMed]

16. Morin, Lucas, Regis Aubry, Yvan Beaussant and Philippe Rochigneux, et al. “Burden of Inpatient Care and Treatments in Terminally-Ill Cancer Patients Results from a Population-Based, Retrospective Study from Administrative Data in France.” J Clin Oncol 35(2015): 9527-9527.

    [Crossref] [Google Scholar]