Abstract
Background/Aim: The general pattern of care regarding the application of prophylactic cranial irradiation (PCI) in small cell lung cancer (SCLC) patients in Germany has not been previously evaluated. This survey was conducted to assess patterns of care. Patients and Methods: Radiation oncology institutions in Germany were surveyed via an anonymous online questionnaire sent by e-mail to member institutions of the German Society for Radiation Oncology. Results: A total of 95 responses were received (29% response rate). Eighty-eight percent of all responders recommended a PCI total dose of 30Gy delivered in 15 daily fractions. Overall, 11 and 38% of the respondents applied PCI simultaneously with chemo- and radiotherapy, respectively. A quarter of respondents offered hippocampal-avoidance PCI and followed their patients with serial brain imaging. Conclusion: PCI with a total dose of 30 Gy in 15 daily fractions, without neuropsychological testing and hippocampus-avoidance, delivered after completion of primary multimodal treatment remains standard in Germany.
Small cell lung cancer (SCLC) is a highly aggressive, recalcitrant and lethal tumour entity which accounts for approximately 15% of lung malignancies (1). Approximately 10% of subjects present at diagnosis with brain metastasis (BM) with a cumulative risk rising to 50% (2) and an up to 80% incidence of BM at autopsy (3).
Brain metastasis is highly detrimentive and associated with poor overall survival. In order to predict survival in limited-staged (LS) SCLC, several prognostic factors such as Karnofsky performance status, nodal lymph involvement and the application of prophylactic cranial irradiation (PCI) have been identified (4-6). These factors are helpful in understanding a patient's prognosis and can guide physicians to improve treatment allocation and optimize individual treatment planning.
In 1999, a landmark meta-analysis by Aupérin and colleagues (7) demonstrated an overall survival benefit from the delivery of PCI in SCLC complete responders following induction therapy, which decreased cumulative incidence of BM.
The National Comprehensive Cancer Network (NCCN) guidelines recommend PCI in patients with LS-SCLC who have a good response to initial therapy based on the above-mentioned meta-analysis and trials which did not incorporate routine brain imaging (8). As such, critics have repeatedly questioned its effect on overall survival (OS) and its potential detrimental effect on neuro-cognition in the current era of universal availability of modern MRI neuroimaging.
Since the recent publication of the highly-anticipated Japanese study by Takahashi et al. (9), universal delivery of PCI in extensive staged (ES) SCLC has come under increased scrutiny as PCI did not result in prolonged OS.
We conceptualised an online survey to gauge common clinical practice in German radiation oncology facilities regarding the management of SCLC patients with emphasis in the current study on PCI. Herein the results of this survey are presented.
Patients and Methods
Survey design. The anonymous survey was designed with the online professional survey tool LimeSurvey licensed for use by the Ludwig-Maximilians-University of Munich. The survey contained 29 multiple-choice questions regarding respondent characteristics, demographics and PCI practices. An e-mail message with a link to the 29-item web-based questionnaire was sent to approximately 348 listed radiation oncology facilities compiled through the Deutsche Gesellschaft fuer Radioonkologie e.V. (DEGRO) directory which is the official German society for radiation oncologists. In order to obtain representative results, the questionnaire was deliberately sent out to the given institutional e-mail addresses and not to individual DEGRO members, hence targeting more experienced radiation oncologists. The survey invitation contained rationale, instructions on participation and contact information on participation. The invitations were initially sent out on August 9, 2017 with a reminder ensuing on August 22, 2017 to maximise response rate.
Respondents were instructed to select answers from a multiple-choice questionnaire closest to their own clinical practice and were characterised by age, geographical location, years of experience and number of patients with lung cancer treated annually.
Responses were collected from August to October 2017. Twenty-three of the 29 questions were evaluated, and all complete responses were deemed eligible for analysis using descriptive statistics. Ethics approval for a pattern of care study comprising an online questionnaire was not applicable.
Our findings were based on the responses of 64 experienced practicing radiation oncologists.
Ethics statement. Ethics approval for a pattern of care study comprising an online questionnaire was not applicable. Furthermore, this registry does not meet the WHO definition of a clinical trial and is considered exempt from clinicaltrials.gov requirements.
Results
The Survey was sent to 348 institutional e-mail addresses. Ten failed/undeliverable automatic responses and 13 automatized out of office responses ensued.
A total of 95 responses were received (29% response rate). Of those 64 were completed and returned, and hence were eligible for further evaluation. Sixty-one percent of responders were between the ages of 50-59 years and 88% had over 15 years of experience in the management of lung malignancies. Overall, 47% of responders treated more than 15 SCLC patients per year.
The characteristics of the 64 complete respondents after exclusion of automatized/failed responses (19.7% response rate) are summarised in Table I.
The survey included a substantial number of tertiary referral hospitals in particular academic medical centers (high-volume centers) treating 50-100 lung cancer patients per annum. SCLC patients were comprising 10-20% of lung cancer patients; approximately 15-30 patients per year. Most institutions administered chemoradiotherapy (CRT) completely in-house and prior to initial treatment, 81% of respondents discussed >90% of cases in multidisciplinary tumour boards. All patients were treated according to the national guidelines of the German Respiratory Society e.V. (DGP e.V.), the German Cancer Society e.V. (DKG e.V.) (10), the international guidelines of the European society for medical oncology (ESMO) (11) and NCCN guidelines (12).
The absolute majority of respondents, in the order of 97%, recommended PCI in LS-SCLC and 67% in ES-SCLC (Table II). 88% recommended a PCI total dose of 30Gy delivered in 15 daily fractions. 11 and 38% of respondents applied PCI simultaneously with chemo- and radiotherapy, respectively. A quarter of respondents offered hippocampal-avoidance PCI with a similar number of fractions following their patients with serial cranial MRI.
Overall, 56 (88%) institutions do not routinely perform neurocognitive testing prior to primary treatment and prior to PCI.
Following PCI, a quarter of respondents follow their patients regularly with serial brain imaging (cranial MRI). On onset of metachronous brain failure, 38 (59%) respondents recommended whole brain radiotherapy (WBRT) or sterotactic radiosurgery (SRS) depending on the number of BM and 28% recommended SRS irrespective of the number of BM.
Discussion
Since its initial proposal in the 1970s, PCI has been the topic of recurring discussion as experts have serially debated its pros and cons in particular in relation to potential neurologic sequelae. Historically, Cmelak and colleagues (13) first published a large survey with a total of 1231 responders (13.4% response rate), including 628 (51%) radiation oncologists, 587 (48%) medical oncologists, 8 (0.6%) surgical oncologists, and 8 (0.6%) from other oncology specialties. Of respondents, 74% recommended PCI in LS-SCLC, including 65% of medical oncologists and 82% of radiation oncologists (p=0.001). Only 30% of respondents recommended PCI for ES-SCLC patients (p=0.001). Medical oncologists believed more often than radiation oncologists that PCI causes late neurological sequelae (95% vs. 84%, p<0.05). However, overall only 1.5% routinely obtained neuropsychiatric testing in PCI patients.
LS-SCLC. Since the above-mentioned meta-analysis, PCI has been standard-of-care for patients with LS-SCLC who had good response to initial treatment, as a small but significant overall survival benefit from the administration of PCI in complete responders led to a 5.4% increase in 3-year survival. Moreover, PCI increased the rate of disease-free survival of recurrence or death, 0.75; 95%CI=0.65-0.86; p<0.001) and decreased the cumulative incidence of BM (RR=0.46; 95%CI=0.38-0.57; p<0.001) (14). However, the criticism, levelled against the meta-analysis, was that it included studies in an era in which consequent neuroimaging was not mandated (pre-MRI era), and hence the role of PCI in the current age of ubiquitous availability of brain imaging might be subject to re-evaluation.
In 2016, Shahi et al. (15) published a 35-item survey on the management of SCLC by Canadian radiation oncologists and revealed that PCI was universally offered to therapy responders with LS-SCLC (100%). Performance status, baseline cognition, and pre-PCI brain imaging were important patient factors evaluated prior to PCI. However, the role of neuropsychological testing and hippocampus avoidance was not evaluated in this survey.
In a recent survey of US radiation oncologists, 98% recommended PCI for patients with LS-SCLC and 35% recommended memantine for patients undergoing PCI. Recommendation of memantine was associated with fewer years of post-residency training (16).
ES-SCLC. The delivery of PCI in ES-SCLC is the subject of even more intensive debate. Since the EORTC study by Slotman and colleagues (17) which demonstrated a) a reduction in the risk of BM in the PCI group (HR=0.27; 95% confidence interval (CI), 0.16 to 0.44; p<0.001), b) cumulative risk of brain metastases within 1 year of 14.6% in the PCI group (95%CI=8.3-20.9) vs. 40.4% in the control group (95%CI=32.1-48.6), c) 1-year survival rate of 27.1% (95%CI=19.4-35.5) in the PCI group and 13.3% (95%CI=8.1-19.9) in the control group, d) association of PCI with an increase in median disease-free survival from 12.0 weeks to 14.7 weeks and e) an increase in median OS from 5.4 months to 6.7 months from randomisation, PCI has been adopted as a standard of care for ES-SCLC therapy responders. However, a recently published Japanese study has sparked a new debate on the universal delivery of PCI in ES-SCLC treatment responders. The study by Takahashi et al. (9) enrolled, between 2009 and 2013, 224 patients who were randomly assigned (113 to PCI vs. 111 to observation). In the planned interim analysis on June 18, 2013, of the first 163 enrolled patients, Bayesian predictive probability of PCI being superior to observation was 0.011%, resulting in early termination of the study due of futility. In the final analysis, median OS was 11.6 months (95%CI=9.5-13.3) in the PCI group vs. 13.7 months (10.2-16.4) in the observation group (HR 1.27, 95%CI=0.96-1.68; p=0.094). The authors concluded that PCI could be deferred in therapy responders and instead, patients be followed intensively by serial cMRI with adoption of radiotherapy (WBRT vs. SRS) in the salvage setting. The authors iterate that in the EORTC study, only 29% of randomised patients received brain imaging at diagnosis and the number of patients who received brain imaging prior to PCI is not clear. In contrast, comprehensive serial cMRI was mandatory in the Japanese study.
In 2016, survey of 473 U.S. radiation oncologists on their recommendations regarding PCI in ES-SCLC revealed that 98% recommended PCI for therapy responders albeit only half of these recommendations ultimately led to patients actually receiving PCI (18). Half of the respondents followed their patients with brain MRI after completion of PCI. One-third of the respondents prescribed the neuroprotectant N-methyl-D-aspartate receptor antagonist memantine to patients undergoing PCI during the RTOG 0614 trial. This double-blind and placebo-controlled trial which randomised adult patients with BM receiving WBRT to receive placebo or memantine was, as the authors duly noted, technically a negative trial (19). The NCCN guidelines however suggest considering memantine-use in this scenario.
Pattern of care according to current survey. Our survey demonstrated that PCI is universally adopted in LS SCLC. However, in ES SCLC only 67% respondents recommend PCI. The preferred fractionation in Germany irrespective of disease stage was 30/2 Gy fractions, which differs to the universally adopted standard of 25 in 2.5 Gy single fractions. Importantly, the absolute majority of radiation oncologists surveyed do not perform neuropsychological testing at any time-point before, during and/or after the PCI treatment.
A quarter of respondents offer hippocampal-avoidance (HA-) PCI; probably owing to the growing body of evidence on this topic (20, 21) and the currently active studies on this issue (NCT02736916, NCT02906384/ZJCH-HA-PCI, NCT02397733, NCT01780675/M12PHA, NCT02635009/NRG-CC003 trials). In the study by Gondi et al., conformal avoidance of the hippocampus during WBRT was associated with preservation of memory and QOL compared with a historical patient cohort (20). These findings go along with the prospective study by Redmond et al. (21). The Authors found a potential neurological benefit of hippocampal sparing as well, but with a risk of failures in the spared region. They suggest supporting the enrollment on ongoing cooperative group randomized trials to improve the level of evidence. However, an evaluation of the clinical relevance of hippocampal-sparing PCI without standardized and continuous assessment of patient cognitive functions remains questionable. Additionally, a probability of metastatic risk in the avoidance regions needs further prospective evaluation with prolonged follow-up. Neuroprotectants are not routinely prescribed in Germany in this setting. Currently, however, there are several international prospective studies evaluating neuroprotectants together with PCI.
The strength of the current survey was the substantial number of tertiary referral hospitals in particular high-volume centres treating 50-100 lung cancer patients per annum. Furthermore, 88% respondents of our survey had >15 years of experience.
Acknowledging the limitations of our study, it is conceded that the overall response rate was lower than expected with one fifth of our target population responding. In addition, the survey exclusively targeted only German radiation oncology centres; hence the results may not be representative of practice among radiation oncologists outside Germany. Furthermore, selection bias was also inevitable as more interested participants were more likely to respond to the survey. Nonetheless, to the best of our knowledge, this is the premier German survey pertaining to this important issue.
Footnotes
Conflicts of Interest
All Authors have declared that there are no conflicts of interest with regard to this work.
- Received July 30, 2018.
- Revision received August 12, 2018.
- Accepted August 16, 2018.
- Copyright© 2018, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved