Patterns of Care for Craniopharyngioma: Survey of members of the American Association of Neurological Surgeons

Abstract

Background/Significance

Initial therapy for craniopharyngioma remains controversial. Population-based datasets indicate that traditional algorithms (GTR versus STR +/− XRT) are often not employed. We investigated neurosurgical practice patterns.

Methods

A ten-question survey was electronically distributed to members of the American Association of Neurological Surgeons. Responses were analyzed using standard statistical techniques.

Results

One hundred-two responses were collected, with a median 25 craniopharyngiomas managed per respondent. 36% estimated their practice included ≥75% pediatrics and 61% had an academic practice. 36% would recommend observation or radiation therapy for a suspected craniopharyngioma in the absence of a tissue diagnosis, with 46% of these indicating this recommendation in ≥10% of cases. Following STR, 35% always recommend XRT and 59% recommend it in over half of cases. However, following STR or biopsy alone, 18% and 11% never recommend XRT. There was no association between type of practice (i.e. academic or ≥75% pediatrics) and practice patterns.

Conclusions

This survey verifies that deviation from established algorithms is common, underscoring the clinical complexity of these patients and recent secondary data analyses This should influence clinical researchers to investigate outcomes for patients treated using alternative methods. This will lend insight into appropriate treatment options and contribute to quality of life outcomes studies for craniopharyngioma.

Introduction

Initial therapy for both adult and pediatric patients with craniopharyngioma has been the subject of considerable controversy. While the most commonly described algorithm centers around the choice between Gross Total Resection (GTR) and Subtotal Resection (STR) augmented by external beam radiation (RT), multiple groups promote an individually tailored treatment approach [1–6]. Two recent analyses of the National Cancer Institute’s Surveillance, Epidemiology and End Results (SEER) datasets indicate that a significant portion of patients in the United States are not initially treated with GTR or STR + RT [7,8]. While alternative therapeutic regimens, such as intracystic therapy or progression contingent radiation are well established, these data prompted us to further explore the initial practice patterns of neurosurgeons using a brief survey. We intended to survey as broad a population of practicing neurosurgeons as possible, with the knowledge that this would predispose the instrument to a low response rate. We hypothesized that neurosurgeons commonly elect initial treatment strategies other than gross total surgical resection or subtotal resection followed by radiation therapy. Clarification of initial practice patterns not only offers the potential to detail the application of less commonly described approaches, but also informs efforts to design clinical outcomes studies for patients with craniopharyngioma.

Methods

A ten question survey was designed then reviewed by leadership of the Joint Sections of the American Association of Neurological Surgeons (AANS) and Congress of Neurological Surgeons on Tumors and Pediatric Neurosurgery. The survey was electronically distributed through 2 electronic AANS Special Announcements to 2974 AANS members in June and July, 2012 (see Appendix #1). Results were collected through Zoomerang (SurveyMonkey, Palo Alto, CA.). Statistical analysis was performed using SAS^® 9.3 (SAS^® Institute, Cary, NC). Frequencies, proportions and measures of central tendency were used to describe study results and chi square analysis was used to evaluate bivariable associations. Statistical significance was set at a value of p <0.05.

Results

Respondent Characteristics

Responses to 102 surveys (3.4%) were received. The estimated number of patients with craniopharyngioma that each respondent had assessed during her/his career ranged from 1–250 (median = 25), for an estimated total of 2550 cases. Thirty-seven (36.3%) respondents reported their practice to include ≥75% pediatric patients and 62 (60.8%) described their practice as academic. Thirty-two respondents (31.4%) completed neurosurgical training between 1977 and 1989, 37 (36.3%) between 1990 and 1999, and 33 (32.4%) between 2000 and 2008.

Respondents who described their practices as academic were more likely to report evaluating more than 25 patients with craniopharyngioma, compared to those in private practice (54.8% vs. 22.5%, p=.001).

Initial Treatment

In response to the question “Have you ever you ever recommended observation or treatment without a tissue diagnosis for a patient presenting with a newly diagnosed lesion considered likely to be a craniopharyngioma?” 37 (36.3%) respondents responded in the affirmative. They reported this recommendation in 0.2–80% of cases (median = 5%), with 17 (16.8% of total sample) respondents reporting this recommendation in at least 10% of cases. Questions 7–9 (see Addendum #1) queried how often respondents recommend radiation therapy immediately following surgery (GTR, STR, and Bx, respectively). One hundred respondents (99.0%) do not recommend RT in the majority of cases in which GTR is achieved. Following STR, 18% never recommend RT, 41% recommend it in half or fewer of the cases, and 35% always recommend RT. Following biopsy alone, 56% always recommend post-operative RT and 71% recommend it in ≥90% of cases. Following biopsy alone, 11/99 (11.1%) never recommend RT.

There were no statistically significant differences in the initial treatment patterns of respondents based on the demographic parameters assessed: experience (based on date of training or number of craniopharyngiomas evaluated), practice type, and proportion of practice that is composed of pediatric patients (Tables 1&2). Regarding the recommendation for treatment or observation without tissue biopsy, 66.1% of those in academic practice have never done so and 60.0% of those in private practice have never done so (p=0.53).

Table 1.

Reported likelihood of recommending XRT following STR (n = 100)

	Patients
	≤50%	>50%	p value
Practice experience
Year of training completion
1977–1989	13	19
1990–1999	19	17
2000–2008	9	23	0.12
Number craniopharyngiomas evaluated
0–25	27	31
30–250	14	28	0.19
Practice character
Proportion of pediatric patients
≥75%	11	25
<75%	30	34	0.11
Practice setting
Academic	22	40
Private	19	19	0.15

Table 2.

Reported likelihood of recommending XRT following biopsy (n = 99)

	Patients		p value
	≤50%	>50%
Practice experience
Year of training completion
1977–1989	6	26
1990–1999	10	26
2000–2008	4	27	0.31
Number craniopharyngiomas evaluated
0–25	14	43
30–250	6	36	0.21
Practice character
Proportion of pediatric patients
≤75%	6	30
<75%	14	49	0.51
Practice setting
Academic	12	50
Private	8	29	0.79

Discussion

Craniopharyngioma is a tumor that is well known for the options that exist regarding initial surgical management. While considerable literature compares the merits and limitations of initial GTR versus STR and external beam radiation as the mainstays of initial therapy [5,9–19], alternative algorithms including intracystic therapies [20–26], progression contingent radiation [27], and stereotactic radiosurgery [28–31] have also been well described through both primary and secondary data analyses [7,8,32,33]. This survey intended to expand on recent secondary data analyses by assessing how commonly alternative algorithms are employed by contemporary neurosurgeons. We found that many patients are managed with algorithms that differ from the pattern of GTR or STR + XRT. While not statistically significant, there was a trend towards neurosurgeons who completed training between 2000–2008 to employ XRT following STR or Bx at greater rates than their more senior colleagues (tables 1&2). This may reflect a practice trend towards greater use of XRT, but further studies are required to more clearly address this question. We were surprised that 18% and 11% of respondents never recommend XRT immediately following STR or tumor biopsy, respectively. Additionally, 36.3% (37/102) of respondents indicated that they had treated or observed a patient with a presumptive craniopharyngioma without a biopsy. Although not a direct comparison, this is consistent with the rates of 12.7–20.8% of patients identified from the SEER datasets as receiving neither surgery nor radiation therapy at the time of diagnosis with craniopharyngioma [7,8]. However, such practice patterns lie in distinction to reports from multiple high volume centers, in which all reported patients were treated with surgery, with or without postoperative radiation therapy [5,14,16,19]. Furthermore, recent systematic reviews [10–11], while they provide excellent insight regarding comparative outcomes, focus on patients who were treated operatively, thereby excluding what may be a substantial subset of patients.

These results are significant because they reinforce the concept that therapy for patients with craniopharyngioma is often individually tailored [2–6]. Furthermore, variability in initial treatment algorithms must be incorporated into the design and interpretation of high quality clinical outcomes studies for patients with craniopharyngioma. While recent analyses of large retrospective datasets have associated post-treatment morbidity/mortality with factors such as surgical experience [34], extent of resection [34], age at presentation [18], tumor size [19] and hydrocephalus [18,19], these conclusions are limited by the bias associated with their datasources, which include data collected over a wide time range [18], pooled data [34], and a study design focused on the difference in outcomes between primary and recurrent craniopharyngioma treatment [19].

While randomized clinical trials for craniopharyngioma are logistically challenging [35], pragmatically designed multicenter registries offer the opportunity for both retrospective data analysis using consistent definitions and prospective data collection. The trinational KRANIOPHARYNGEOM 2000 study collected observational data from 46 centers [36]. Among their findings, the authors identified low 3-year event free survival rates in both the completely (EFS 0.64) and incompletely (EFS 0.31) resected populations. The subsequent study, KRANIOPHARYMGEOM 2007, includes randomization of a subset of patients in an effort to ascertain the optimal timing of radiation therapy following incomplete surgical resection in children over 5 years of age [37]. Currently, there is no equivalent multicenter study ongoing in North America.

Despite the limited scope of this survey instrument and recognition that our highly diverse surveyed population would likely yield a low response rate, the response rate of 3.4% was suboptimal. As a result, response bias may lead to conclusions that are not generalizable. This point is demonstrated by a bias in the respondents towards both academic (60.8%) and pediatric practice (36.3%). Nevertheless, the data represent the practice patterns of over 100 neurosurgeons, and represents a large range of experience, based on the years in practice and number of craniopharyngiomas assessed. Additionally, with nearly 40% of the respondents practicing outside an academic environment, we feel that these data do represent both the academic and private practice environments. A larger sample would be unlikely to refute the conclusion that a variety of approaches are used in the initial treatment of patients with craniopharyngioma. A second limitation of this work resulted from our effort to simplify the questionnaire as much as possible. This presented challenges to the respondents by limiting their capacity to answer with the optimal level of detail. Specific concerns resulted from our choices to exclude questions regarding the use of intracavitary therapy and to not rigidly define a timeframe for “immediate post-operative” radiation. The latter decision was made to allow respondents the freedom to express their intent regarding the timing of therapeutic recommendations. Future clinical outcomes studies will allow for more detailed responses.

Conclusions

Using a streamlined survey instrument that gathered data from members of the American Association of Neurological Surgeons, we verified recent study results indicating that many patients with craniopharyngioma receive initial treatment that is tailored to the clinical situation, often deviating from the most common algorithm: GTR versus STR and Radiation Therapy [7,8]. This finding verifies that deviation from this algorithm is common and should influence clinical researchers to investigate and document the outcomes of patients who are treated using alternative methods. This will not only lend insight into appropriate treatment options, but will contribute to the design of prospective quality of life outcomes studies for patients with craniopharyngioma. Given the rarity of this tumor, pragmatic trial designs that accommodate less restrictive treatment algorithms will be necessary in order to complete such studies.