Introduction

Radiotherapy is one of the pillars in glioblastoma management. As glioblastoma is notorious for its extensive tumor infiltration, the 2023 ESTRO-EANO guidelines [1] recommend a 15 mm clinical target volume (CTV)-margin, adjusted for anatomical barriers, around the gross tumor volume (GTV). In the majority of cases, this expansion adequately encompasses tumor infiltration. However, these large target volumes can also encompass considerable amounts of healthy brain tissue, which increases the risk of radiotherapy-related adverse events [2], [3], [4], [5]. Identifying opportunities to safely reduce the CTV-margin, without compromising tumor control, may allow radiotherapy to achieve similar therapeutic efficacy while preserving the quality of life.

Several studies have highlighted the potential of various smaller CTV-margins [6], [7], [8], [9], showing a pattern of failure consistent with that of cohorts treated with a conventional CTV-margin. For instance, Paulsson et al. [8] retrospectively compared the pattern of failure in 161 patients treated with different CTV-margins and found no statistically significant differences in pattern of failure: In-field recurrence was observed in 79%, 77%, and 87% of patients treated with CTV-margins of 5 mm, 10 mm, and 15–20 mm, respectively. The comparable pattern of failure implies that tumor infiltration was as effectively targeted when patients were irradiated with reduced versus conventional CTV-margins.

Pattern of failure analysis typically examines the spatial relationship between the recurrence volume and the high-dose radiation treatment field (e.g., the 95% isodose line of a radiotherapy plan). By comparing the pattern of failure from conventional radiotherapy plans with those from (theoretical) radiotherapy plans employing a reduced CTV-margin, one can infer the potential impact of CTV-margin reduction. This approach to pattern of failure analysis has long guided efforts to define more optimal CTV-margins for glioblastoma. Yet, this strategy has its limitations, as the observed pattern of failure inherently depends on the initially-chosen CTV-margin and does not provide information on the potential of other CTV-margins. Exploring the distances between the recurrence volume and GTV may better reveal the spectrum of potentially effective CTV-margins. Moreover, recurrence distance analysis may provide a pathway to exploring the concept of personalized CTV-margins. Since the extent of tumor infiltration may vary among patients [10], [11], CTV-margins that are tailored to the individual patient may offer additional value for optimizing target delineation.

In this work, we assessed the distribution of distances between the GTV and recurrence volume in a retrospective cohort of patients with glioblastoma, treated with standard chemoradiation. Furthermore, we performed multiple linear regression modeling to identify variables that were associated with the distance between the GTV and the recurrence volume, and to evaluate the potential for personalized CTV-margins.