Volumetric de-escalation and improved acute toxicity with proton craniospinal irradiation using a vertebral body-sparing technique

Late-onset lymphopenia during radiation is associated with an increased risk of tumor recurrence in newly diagnosed pediatric medulloblastoma

BACKGROUND

Recent data found a correlation between lymphopenia occurring early during craniospinal radiation therapy (RT) and risk of disease recurrence in newly diagnosed childhood medulloblastoma. However, the population included patients who received chemotherapy prior to or during RT. Here, we investigate the effect of lymphopenia during RT in patients with newly diagnosed pediatric medulloblastoma who were chemotherapy-naïve.

PROCEDURE

We analyzed 79 patients with newly diagnosed medulloblastoma (ages 2-21 years) treated between 1997 and 2013 with craniospinal RT. Log-rank tests were used to determine survival differences, and Cox proportional hazards regression was used to assess associations between patient characteristics and lymphopenia with disease recurrence risk.

RESULTS

Eighty-three percent of patients (62/75) had grade ≥3 lymphopenia by RT Week 3, with 95% developing grade ≥3 lymphopenia at some point during therapy. There was no difference in incidence of lymphopenia between those who received proton beam RT (93%) versus photon (97%). Twenty-four of 79 (30%) patients developed disease recurrence at an average 27.0 months after diagnosis. There was higher risk of disease recurrence in patients with grade ≥3 lymphopenia during RT Week 4 (log-rank p = .016; Cox p = .03) and Week 5 (log-rank p = .024; Cox p = .032); after adjusting for clinical risk group, only grade ≥3 lymphopenia at Week 4 remained prognostic (Cox p = .04). No correlation was found between risk of tumor recurrence and early lymphopenia (RT Weeks 0-3) or absolute lymphocyte count (ALC) below the median at any time during RT.

CONCLUSIONS

Lymphopenia during RT Weeks 4 and 5 correlates with increased risk of tumor recurrence in pediatric patients with newly diagnosed medulloblastoma.

Risks of Spinal Abnormalities and Growth Impairment After Radiation to the Spine in Childhood Cancer Survivors: A PENTEC Comprehensive Review

PURPOSE

A PENTEC (Pediatric Normal Tissue Effects in the Clinic) review was performed to estimate the dose-volume effects of radiation therapy on spine deformities and growth impairment for patients who underwent radiation therapy as children.

METHODS AND MATERIALS

A systematic literature search was performed to identify published data for spine deformities and growth stunting. Data were extracted from 12 reports of children irradiated to the spine (N = 603 patients). The extracted data were analyzed to find associations between complication risks and the radiation dose (conventional fractionation throughout) as impacted by exposed volumes and age using the mixed-effects logistic regression model. When appropriate, corrections were made for radiation modality, namely orthovoltage beams.

RESULTS

In the regression analysis, the association between vertebral dose and scoliosis rate was highly significant (P < .001). Additionally, young age at time of radiation was highly predictive of adverse outcomes. Clinically significant scoliosis can occur with doses ≥15 Gy to vertebrae during infancy (<2 years of age). For children irradiated at 2 to 6 years of age, overall scoliosis rates of any grade were >30% with doses >20 Gy; grade 2 or higher scoliosis was correlated with doses ≥30 Gy. Children >6 years of age remain at risk for scoliosis with doses >30 Gy; however, most cases will be mild. There are limited data regarding the effect of dose gradients across the spine on degree of scoliosis. The risk of clinically meaningful height loss was minimal when irradiating small volumes of the spine up to 20 Gy (eg, flank irradiation), except in infants who are more vulnerable to lower doses. Growth stunting was more frequent when larger segments of the spine (eg, the entire spine or craniospinal irradiation) were irradiated before puberty to doses >20 Gy. The effect was modest when patients were irradiated after puberty to doses >20 Gy.

CONCLUSIONS

To reduce the risk of kyphoscoliosis and growth impairment, the dose to the spine should be kept to <20 Gy for children <6 years of age and to <10 to 15 Gy in infants. The number of vertebral bodies irradiated and dose gradients across the spine should also be limited when possible.

Late Effects of Craniospinal Irradiation Using Electron Spinal Fields for Pediatric Patients With Cancer

PURPOSE

For children, craniospinal irradiation (CSI) with photons is associated with significant toxic effects. The use of electrons for spinal fields is hypothesized to spare anterior structures but the long-term effects remain uncertain. We studied late effects of CSI using electrons for spinal radiation therapy (RT).

METHODS AND MATERIALS

Records of 84 consecutive patients treated with CSI using electrons for the spine at a single institution between 1983 and 2014 were reviewed. Median age at RT was 5 (range, 1-14) years. The most common histologies were medulloblastoma/primitive neuroectodermal tumor (59%) and ependymoma (8%). The median prescribed dose to the entire spine was 30 Gy (range, 6-45). A subset of 48 (57%) patients aged 2 to 14 at RT with clinical follow-up for ≥5 years was analyzed for late effects. Height z scores adjusted for age before and after CSI were assessed using stature-for-age charts and compared with a t test.

RESULTS

At median follow-up of 19 years (range, 0-38 years), the median survival was 22 years (95% confidence interval, 12-28 years) after RT, with 47 patients (56%) alive at last follow-up. On subset analysis for late effects, 19 (40%) patients developed hypothyroidism and 5 (10%) developed secondary malignancies. Other complications reported were esophageal stricture and periaortic hemorrhage in 1 and restrictive pulmonary disease in 1 patient. Median height z score before treatment was -0.4 (36th percentile; interquartile range, -1.0 to 0.0) and at last follow-up was -2.2 (first percentile; interquartile range, -3.1 to -1.6; P < .001). Of 44 patients with spinal curvature assessments, 15 (34%) had scoliosis with median Cobb angle 15° (range, 10°-35°) and 1 (2%) required surgery.

CONCLUSIONS

Frequent musculoskeletal toxic effects and predominantly decreased height were seen with long-term follow-up. Scoliosis and hypothyroidism were each seen in at least one-third of long-term survivors. However, clinically evident esophageal, pulmonary, and cardiac toxic effects were infrequent.

Proton and Carbon Ion Radiation Therapy Decreased Severe Lymphopenia by Reducing Thoracic Vertebra and Aortic Doses in Non-small Cell Lung Cancer Versus Intensity Modulated Radiation Therapy

PURPOSE

Lymphopenia is a common adverse effect of radiation therapy (RT). Little is known about the difference in lymphopenia between intensity modulated (photon) radiation therapy (IMRT) and proton and carbon ion radiation therapy (PCIRT). This study aimed to investigate lymphopenia differences between IMRT and PCIRT in non-small cell lung cancer (NSCLC).

METHODS AND MATERIALS

Clinical and dosimetric parameters were collected from 343 patients who received definitive IMRT or PCIRT for NSCLC. Severe lymphopenia (SRL) was defined as an absolute lymphocyte count (ALC) ≤0.5 × 10³ cells/μL. Overall survival (OS) was analyzed using the Kaplan-Meier method. Propensity score matching was performed between the IMRT and PCIRT groups. Least absolute shrinkage and selection operator analysis was used to select appropriate dosimetric parameters. Univariate and multivariate logistic regression analyses were conducted to identify the predictors of SRL.

RESULTS

Compared with the IMRT group, the PCIRT group was less likely to develop SRL (P < .001). Compared with the non-SRL group, the SRL group showed significant association with poorer OS, with a median survival time of 29.2 versus 15.0 months (P = .046). IMRT was an independent risk factor of SRL (P = .004). A lower ALC before RT (P = .030) and larger planning target volume (PTV) (P = .002) were also significant independent risk factors for SRL. Moreover, the majority of dosimetric parameters of organs at risk in PCIRT were lower than those in IMRT (P < .001). Thoracic vertebra V5 (P = .002) and aorta V5 (P = .026) were identified as independent risk predictors of SRL after adding dosimetric parameters to the regression model.

CONCLUSIONS

Compared with IMRT, PCIRT could reduce SRL incidence, possibly by limiting thoracic vertebra and aortic doses, and SRL was associated with poor outcomes in patients with NSCLC.