PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of jradsbrtLink to Publisher's site
 
J Radiosurg SBRT. 2016; 4(1): 53–60.
PMCID: PMC5658837

Predicting treatment related imaging changes (TRICs) after radiosurgery for brain metastases using treatment dose and conformality metrics

Abstract

Purpose

Treatment-related imaging changes (TRICs) after stereotactic radiosurgery (SRS) involves the benign transient enlargement of radiographic lesions after treatment. Identifying the radiation dose volumes and conformality metrics associated with TRICs for different post-treatment periods would be helpful and improve clinical decision making.

Methods

367 metastases in 113 patients were treated using Gamma Knife SRS between 1/1/2007-12/31/2009. Each metastasis was measured at each imaging follow-up to detect TRICs (defined as ≥ 20% increase in volume). Fluctuations in small volume lesions (less than 108 mm3) were ignored given widely variable conformity indices (CI) for small volumes. The Karolinska Adverse Radiation Effect (KARE) factor, Paddick’s CI, Shaw’s CI, tumor volume (TV), 10 Gy (V10) and 12 Gy (V12) volumes, and prescription isodose volume (PIV) were calculated.

Results

From 0-6 months, all measures correlated with the incidence of TRICs (p<.001), except KARE, which was inversely correlated. During the 6-12 month period all measures except KARE were still correlated. Beyond 12 months, no correlation was found between any of the measures and the development of TRICs.

Conclusions

All metrics except KARE were associated with TRICs from 0-12 months only. Additional patient and treatment factors may become dominant at greater times after SRS.

Keywords: radiosurgery, SRS, conformality metrics, TRICs, KARE, brain metastases

1. INTRODUCTION

As the use of stereotactic radiosurgery (SRS) for brain metastases increases and patient survival lengthens, the incidence of post-SRS adverse radiation effects (ARE) is also likely to rise. An understanding of the factors that contribute to ARE development and how these factors are measured is key to minimizing its incidence. We have previously reported that ARE can be clinically silent but can still be problematic in the form of imaging-based reporting of lesional regrowth that ultimately is shown to be benign and transient [1]. We hypothesized that increasing treatment conformality thereby decreasing normal brain dose should be correlated with a decrease in post-SRS treatment related imaging changes (TRICs). All previously reported indices for measurement of SRS treatment conformality, including the Shaw conformity index [2], the Paddick conformity index [3,4], the Karolinska Adverse Radiation Effect (KARE) factor [5], as well as metrics such as the tumor volume (TV), prescription isodose volume (PIV), and brain tissue volume receiving 10 Gy [6] and 12 Gy [7] (see Table 1), were therefore analyzed and compared for ability to predict TRICs in this single institution prospective study.

Table 1
Description of tested treatment metrics.

2. METHODS

425 metastases in 113 patients were treated using Gamma Knife SRS at our center between 1/1/2007-12/31/2009. Plans were retrospectively reviewed by a single individual and 10 Gy and 12 Gy volumes were determined by the GammaPlan software. Lesions were excluded if 2 lesions were within the same dose calculation matrix or if the 10 Gy isodose surfaces for two lesions intersected. 367 analyzable lesions remained. Follow-up gadolinium-enhanced T1 weighted volumetric MRI scans were performed every 6-12 weeks for all patients until death. At the time of treatment and at each radiographic time point thereafter, lesions were measured in the maximum anterior posterior (AP), right to left (RL), and cranio-caudal (CC) directions by a single individual. Lesion volume at each time-point was calculated using the formula (AP x RL x CC)/2 which very closely approximates the true tumor volume assuming a spherical lesion (4/3)*π*r3. Incidence and timing of any post-SRS lesion re-growth was recorded. Based on our prior work, re-growth beyond a 20% increase in volume was defined as TRIC [1]. For this study, we ignored all fluctuations in lesions with a volume less than 108 mm3 (approximately 6 mm diameter), as small volume changes in these lesions can result in a relatively large apparent percent change but may result purely from discrepancies in imaging technique (slice thickness, timing of gadolinium bolus relative to scan time, and MRI cut variability) or measurement.

For every lesion, the initial TV, PIV, 10 and 12 Gy volumes, Paddick’s CI, Shaw’s CI, and KARE factor (V10/TV) were calculated, and the timing of TRIC development was recorded. Univariate and multivariate logistic regressions were then used to evaluate the ability of each metric to predict TRICs. Age, sex, histology, prescription dose and use of WBRT were also analyzed for influence on TRIC incidence. The entire population of lesions was initially analyzed. Secondary analyses were performed on each quartile of lesion size and on a subset of lesions > 1 cm3 in volume. Finally, a cohort of patients matched to the cohort of Lippitz et al. by survival time (> 9 months post treatment), lesion size (> 0.5 cm3), and V10 (3.4 – 42 cm3) was analyzed. In order to compare the results to the analysis of ARE by Lippitz et al., only TRICs occurring > 6 months post treatment were considered for this subset analysis.

3. RESULTS

3.1. Incidence and Timing of TRIC development

In 367 metastases, the overall rate of TRICs was 18.8%. This is somewhat lower than our previously reported data [1], potentially due to the addition of exclusion criteria for lesions in close proximity with overlapping 10 Gy volumes and ignoring fluctuations in very small lesions, as described above. There was no correlation between TRIC development and age, gender, histology, marginal prescription dose, or use of WBRT on univariate analysis. Increasing survival duration, however, was correlated with an increased likelihood of developing TRICs.

The majority of TRICs for the whole patient group occurred within 6 months of SRS (11.4%), followed by TRICs occurring between 6-12 months (6.2%) and then > 12 months after SRS (1.1%) (Table 2). In patients surviving > 9 months with lesions measuring greater than 0.5 cm3 (as reported by Lippitz et al.), timing of TRICs remained highest within 6 months. However, for the subset of patients surviving > 9 months with lesions measuring less than 0.5 cm3, the peak TRIC incidence occurred between 6-12 months. Smaller lesions thus had both delayed occurrence and decreased overall likelihood of TRICs.

Table 2
Timing of TRIC incidence stratified by survival.

3.2. Correlations between treatment metrics and incidence of TRICs

All metrics correlated as expected with incidence of TRICs, except KARE, which we found to be inversely correlated with TRICs (Table 3). Median KARE indices were smaller for those patients with a TRIC (6.38) compared to those without (8.68, p<.001). Knowing that smaller lesions and treatment volumes are less likely to result in a TRIC, we assumed that the inverse correlation was due to limitation of collimator size causing higher KARE values in such lesions. We then arbitrarily analyzed the subset of lesions with a volume greater than 1.0 cm3 (n=118) and found that only tumor volume and PIV were significantly associated with TRIC development.

Table 3
P-Value results from univariate and multivariate analyses showing correlation between treatment metrics and TRIC outcome.

During the first 6 months after SRS, the more traditional measures (TV, 10 Gy and 12 Gy volume, PIV, Paddick CI, and Shaw’s CI) correlated with TRICs. A smaller correlation between these measures and TRICs was found beyond 6 months. Conversely, within 6 months after SRS, KARE was inversely correlated with TRICs, but no longer correlated with TRIC development beyond 6 months after SRS. Beyond 12 months, no correlation was found between any of the measures and the development of TRICs.

3.3. KARE specific subset

Given the unexpected finding of an inverse correlation with the KARE index, we investigated the subset of patients in our cohort that would be eligible for direct comparison with Lippitz et al. In that study, patients had to survive longer than 9 months with a metastasis of greater than 10 mm diameter (equivalent to volume of > 0.5 cm3 assuming spherical). We found 74 eligible metastases. Median survival in this patient subset was 15 months, compared to 21.2 months reported by Lippitz et al. Similar to the previous study, V10 clearly correlated with the incidence of TRIC development. Our incidence of TRICs beyond 6 months after SRS in each V10 size category also closely matches ARE incidence in Lippitz et al (Table 4). However, when looking specifically at intermediate V10 volumes (3.4 – 42 cm3), our cohort did not have a decreased rate of TRICs for treatment plans with KARE < 4.09. In fact, the rate of TRICs was slightly worse in this lower KARE factor range (Table 5), and the KARE factor did not correlate with overall TRIC development in the entire subset or any quartile within the subset.

Table 4
Risk of ARE and TRIC according to V10 seen at the Karolinska Institute and the current study population.
Table 5
Risk of ARE and TRIC in lesions with V10 between 3.4 cm3 and 42 cm3 stratified by KARE index < 4.09 vs. > 4.09 seen at the Karolinska Institute and the current study.

3.4. Illustrative examples of TRIC variation

Despite the statistical association between treatment metrics and TRICs, treatment metrics did not align perfectly with TRIC incidence, and TRICs remained unpredictable. Two illustrative examples are given in Figure 1, where within the same patient different lesions with identical calculated treatment metrics showed different responses. In addition, we found that late TRICs often occurred with initiation of systemic therapy independent of any treatment metric. Two cases are shown in Figures 2 & 3 where initiation of either a small molecule inhibitor of anaplastic lymphoma kinase (ALK), crizotinib (Figure 2), or an immune modulating anti-PD1 antibody, nivolumab (Figure 3), induced dramatic and symptomatic TRICs after long periods of durable lesion stability.

Figure 1
Two similar brain metastases in a patient with breast cancer (1) and another patient with melanoma (2). Lesions in each respective patient were in similar locations in the brain and were treated on the same date with identical prescription doses. The ...
Figure 2
28 year old female with ALK translocated lung cancer was treated with SRS to numerous brain lesions. Post treatment radiographic time course is shown for 3 of the treated lesions. Robust symptomatic and unexpected radionecrosis was seen in two of the ...
Figure 3
40 year old female with metastatic melanoma was treated with SRS to 32 brain lesions. Post treatment radiographic time course is shown for 8 of the treated lesions. While 3 lesions presented with radionecrosis needing intervention possibly associated ...

4. Discussion

We conclude from this study that, as in previously reported studies, many treatment indices calculated based on lesion size and volume of irradiated brain are correlated with risk of TRIC development. However, when the data was stratified by time after SRS, it became evident that this correlation was most significant in the first 6 months after SRS, less significant in the 6-12 months after SRS, and then no longer correlated beyond 12 months after SRS. This may explain why previous studies have not been able to correlate treatment metrics with risk of ARE.

Although TRIC incidence was highest in the first 6 months after SRS, nearly 40% of the TRICs we observed occurred after 6 months, clearly after periods of stable disease. While regrowth was presumed to be due to radiation effect for all the lesions studied based on their clinical scenario, it is possible that without histopathologic confirmation regrowth for some of the lesions could be due to slowly growing tumor recurrence that was not ultimately recognized if the patient died of progressive systemic disease.

This study was not able to validate the predictive value of the KARE factor for the development of TRICs. There may be a variety of reasons why our results have differed from the original study. Firstly, there was considerable lack of comparability between cohorts, with only 20% of our metastases (74/367) meeting inclusion criteria used by Lippitz et al. In addition, melanoma patients predominated in our cohort, while the Karolinska cohort reported by Lippitz et al. was predominantly lung cancer patients (B. Lippitz, oral communication, May 2014). Volume measurements methods were also different between the 2 studies, and inter-institutional treatment planning variations such as PTV margin around the tumor and number of shots per lesion, etc. may also explain the differences in results. Finally, 65% of our metastases (238 / 367) had a TV <0.5 cm3, and lesions with small TV will tend to have high KARE indices because even the smallest collimator size (4 mm) for the Gamma Knife will not permit highly conformal treatments for small TVs. As TV appears to be a dominant predictor of TRICs, this may explain the inversion correlation observed between the KARE factor and TRICs in our original analysis. However, even after eliminating these smaller lesions and analyzing only those lesions that met the inclusion criteria used by Lippitz et al., we were still unable to validate the use of KARE for predicting TRICs in lesions with V10 from 3.4-42 cm3 or any quartile of lesion size.

In addition to SRS treatment metrics, other patient factors may need to be incorporated into risk modeling. Specific patient and tumor genetics may play a role in SRS outcome and thus likelihood of TRICs. It has been previously reported that lung cancer patients with ALK amplifications or EGFR mutations have extremely high rates of lesional control in the brain after SRS [10]. In addition, as use of more novel targeted agents have emerged, we have seen profound systemic influences possibly inducing imaging abnormalities in the brain after long periods of stability. The phenomenon of “Radiation Recall” has been observed and described previously. It is mostly seen as an acute inflammatory reaction confined to previously irradiated areas that can be triggered when chemotherapy agents (most commonly doxorubicin, docetaxel, paclitaxel, gemcitabine and capecitabine) are administered after radiotherapy [11,12,13,14]. Although one review indicated that the time interval between the completion of radiation therapy and the administration of cytotoxic chemotherapy is between 6 and 37 days in patients that develop recall [11], other reports indicate periods within 90 days [15], to 15 or even 25 years after radiotherapy [16,17]. While most documented cases are limited to the skin due to ease and convenience of diagnosis, there are cases of CNS radiation recall in the current literature [14,18]. In addition to classical cytotoxic chemotherapy, many oncologically useful small molecule targeted therapeutics and monoclonal antibodies have shown profound effects either with or after radiotherapy [19,20,21,22,23,24]. Given the relative lack of information regarding the interaction of novel systemic therapies and SRS related TRICs, this area needs further study.

In addition to novel targeted therapy, we believe immunologic therapy can induce TRICs. Residual long term inflammation and necrosis in brain tissue as well as permanent local disruptions in the vasculature can be seen on histopathology after SRS [25]. The appearance of these inflammatory and necrotic changes are plausibly influenced by agents that can influence immune response. For example, anti-PD-1 therapy (nivolumab) could theoretically enhance immune activity and surveillance in a location harboring chronic inflammation and possibly tumor cell antigens (viable or non-viable), leading to an inflammatory response resulting in a TRIC, as in our patient (see Figure 3). Similarly, ipilimumab, through targeting CTLA-4, an inhibitor of cytotoxic T cells, can produce a similar response on imaging.

The potential for immune agents to induce TRICs is also consistent with other prior work that indicates a more intense inflammatory response for controlled tumors vs. poorly controlled tumors [26,27,25]. For neoplasms well-controlled by SRS, histopathology within the irradiated tumor volume is consistent with a moderate-to-intense inflammatory cell reaction – prominently CD68-positive macrophages and CD3-positive T-lymphocyte populations – that is otherwise missing or sparse in poorly controlled neoplasms with recurrent disease [26]. Given that many of our lesions are “well controlled”, it is plausible that an active and dynamic radiation-induced and immunotherapy enhanced immune response could be responsible for some of the TRICs seen.

In conclusion, our study is unique as it describes SRS related TRICs that can occur after profoundly long periods of stable disease. We found that while the set of metrics available in the literature is useful and should be used for SRS plan evaluation within 12 months of SRS, we were unable to validate their use beyond 12 months. Thus, the influence of systemic therapy on the incidence of TRICs may be overpowering the predictive value of treatment related metrics after 12 months. Due to complex heterogeneity between individual patients, tumor genetics, brain location, and non-targeted and targeted systemic treatments, we need to continue to look for better predictors of long term TRICs after SRS.

Footnotes

Authors’ disclosure of potential conflicts of interest

Drs. Chiang, Knisely, Qian, and Taylor have nothing to disclose.

Dr. Yu reports grants from 21st Century Oncology during the conduct of the study.

Contributed by

Author contributions

Conception and design: Jonathan P. Knisely, Veronica L. Chiang.

Data collection: B. Frazier Taylor.

Data analysis and interpretation: B. Frazier Taylor, Jonathan P. Knisely, Veronica L. Chiang.

Manuscript writing: All authors.

Final approval of manuscript: All authors.

REFERENCES

1. Patel T. R., McHugh B. J., Bi W. L., Minja F. J., Knisely J. P., Chiang V. L. (2011). Acomprehensive review of MR imaging changes following radiosurgery to 500 brain metastases. AJNR Am. J. Neuroradiol., 32, 1885-1892. [PubMed]
2. Shaw E., Kline R., Gillin M., Souhami L., Hirschfeld A., Dinapoli R., Martin L. (1993). RadiationTherapy Oncology Group: radiosurgery quality assurance guidelines. Int. J. Radiat. Oncol. Biol. Phys., 27, 1231-1239. [PubMed]
3. Paddick I. (2000). Asimple scoring ratio to index the conformity of radiosurgical treatment plans. Technical note. J. Neurosurg., 93 Suppl 3, 219-222. [PubMed]
4. Nakamura J. L., Verhey L. J., Verhey L. J., Smith V., Lamborn K. R., Larson D. A., Wara W. M., McDermott M. W., Sneed P. K. (2001). Doseconformity of gamma knife radiosurgery and risk factors for complications. Int. J. Radiat. Oncol. Biol. Phys., 51, 1313-1319. [PubMed]
5. Lippitz Bodo. (2011). ALGORITHM TO AVOID ADVERSE EFFECTS AFTER GAMMA KNIFE RADIOSURGERY OF BRAIN METASTASES. 10th annual International Stereotactic Radiosurgical Society Congress (ISRS): Brain and Body Radiosurgery. 89-89.
6. Voges J., Treuer H., Sturm V., Buchner C., Lehrke R., Kocher M., Staar S., Kuchta J., Muller R. P. (1996). Riskanalysis of linear accelerator radiosurgery. Int. J. Radiat. Oncol. Biol. Phys., 36, 1055-1063. [PubMed]
7. Korytko T., Radivoyevitch T., Colussi V., Wessels B. W., Pillai K., Maciunas R. J., Einstein D. B. (2006). 12 Gy gamma knife radiosurgical volume is a predictor for radiation necrosis in non-AVM intracranial tumors. Int. J. Radiat. Oncol. Biol. Phys., 64, 419-424. [PubMed]
8. Minniti G., Clarke E., Lanzetta G., Osti M. F., Trasimeni G., Bozzao A., Romano A., Enrici R. M. (2011). Stereotacticradiosurgery for brain metastases: analysis of outcome and risk of brain radionecrosis. Radiat. Oncol., 6, 48- [PMC free article] [PubMed]
9. Blonigen B. J., Steinmetz R. D., Levin L., Lamba M. A., Warnick R. E., Breneman J. C. (2010). Irradiatedvolume as a predictor of brain radionecrosis after linear accelerator stereotactic radiosurgery. Int. J. Radiat. Oncol. Biol. Phys., 77, 996-1001. [PubMed]
10. Johung K. L., Yao X., Li F., Yu J. B., Gettinger S. N., Goldberg S., Decker R. H., Hess J. A., Chiang V. L., Contessa J. N. (2013). Aclinical model for identifying radiosensitive tumor genotypes in non-small cell lung cancer. Clin. Cancer Res., 19, 5523-5532. [PubMed]
11. Burris H. A., III, Hurtig J. (2010). Radiationrecall with anticancer agents. Oncologist., 15, 1227-1237. [PMC free article] [PubMed]
12. D’Angio G. J., Farber S., Maddock C. L. (1959). Potentiationof x-ray effects by actinomycin D. Radiology. 73, 175-177. [PubMed]
13. Camidge R., Price A. (2001). Characterizingthe phenomenon of radiation recall dermatitis. Radiother. Oncol., 59, 237-245. [PubMed]
14. Azria D., Magne N., Zouhair A., Castadot P., Culine S., Ychou M., Stupp R., Van Houtte P., Dubois J. B., Ozsahin M. (2005). Radiationrecall: a well recognized but neglected phenomenon. Cancer Treat. Rev., 31, 555-570. [PubMed]
15. Haffty B. G., Vicini F. A., Beitsch P., Quiet C., Keleher A., Garcia D., Snider H., Gittleman M., Zannis V., Kuerer H., Whitacre E., Whitworth P., Fine R., Keisch M. (2008). Timingof Chemotherapy after MammoSite radiation therapy system breast brachytherapy: analysis of the American Society of Breast Surgeons MammoSite breast brachytherapy registry trial. Int. J. Radiat. Oncol. Biol. Phys., 72, 1441-1448. [PubMed]
16. Burdon J., Bell R., Sullivan J., Henderson M. (1978). Adriamycin-induced recall phenomenon 15 years after radiotherapy. JAMA, 239, 931- [PubMed]
17. Barlesi F., Tummino C., Tasei A. M., Astoul P. (2006). Unsuccessfulrechallenge with pemetrexed after a previous radiation recall dermatitis. Lung Cancer. 54, 423-425. [PubMed]
18. Jeter M. D., Janne P. A., Brooks S., Burstein H. J., Wen P., Fuchs C. S., Loeffler J. S., Devlin P. M., Salgia R. (2002). Gemcitabine-induced radiation recall. Int. J. Radiat. Oncol. Biol. Phys., 53, 394-400. [PubMed]
19. Onal C., Abali H., Koc Z., Kara S. (2012). Radiationrecall pneumonitis caused by erlotinib after palliative definitive radiotherapy. Onkologie., 35, 191-194. [PubMed]
20. Arakawa H., Johkoh T., Sakai F., Kusumoto M., Hataji O., Taguchi O. (2011). Exacerbationof radiation fibrosis with erlotinib: another pattern of radiation recall phenomenon. Jpn. J. Radiol., 29, 587-589. [PubMed]
21. Togashi Y., Masago K., Mishima M., Fukudo M., Inui K. (2010). Acase of radiation recall pneumonitis induced by erlotinib, which can be related to high plasma concentration. J. Thorac. Oncol., 5, 924-925. [PubMed]
22. Dauendorffer J. N., Dupuy A. (2009). Radiationrecall dermatitis induced by erlotinib. J. Am. Acad. Dermatol., 61, 1086- [PubMed]
23. Reigneau M., Granel-Brocard F., Geoffrois L., Bauman A. S., Trechot P., Barbaud A., Schmutz J. L. (2013). Efflorescenceof scalp cysts during vemurafenib treatment following brain radiation therapy: a radiation recall dermatitis? Eur. J. Dermatol., 23, 544-545. [PubMed]
24. Anker C. J., Ribas A., Grossmann A. H., Chen X., Narra K. K., Akerley W., Andtbacka R. H., Noyes R. D., Shrieve D. C., Grossmann K. F. (2013). Severeliver and skin toxicity after radiation and vemurafenib in metastatic melanoma. J. Clin. Oncol., 31, e283-e287. [PubMed]
25. Rauch P. J., Park H. S., Knisely J. P., Chiang V. L., Vortmeyer A. O. (2012). Delayedradiation-induced vasculitic leukoencephalopathy. Int. J. Radiat. Oncol. Biol. Phys., 83, 369-375. [PubMed]
26. Szeifert G. T., Atteberry D. S., Kondziolka D., Levivier M., Lunsford L. D. (2006). Cerebralmetastases pathology after radiosurgery: a multicenter study. Cancer. 106, 2672-2681. [PubMed]
27. Szeifert G. T., Salmon I., Rorive S., Massager N., Devriendt D., Simon S., Brotchi J., Levivier M. (2005). Doesgamma knife surgery stimulate cellular immune response to metastatic brain tumors? A histopathological and immunohistochemical study. J. Neurosurg., 102 Suppl, 180-184. [PubMed]

Articles from Journal of Radiosurgery and SBRT are provided here courtesy of Old City Publishing