European Radiology: white matter changes in patients with different prognosis after radiotherapy for nasopharyngeal carcinoma

Nasopharyngeal carcinoma (NPC) is a malignant head and neck malignancy endemic in southern China, especially Guangdong and Guangxi provinces.

Due to the radiation sensitivity of nasopharyngeal carcinoma, radiotherapy (RT) plus adjuvant chemotherapy remains the mainstay of treatment for
nasopharyngeal carcinoma.
However, radiation-induced necrosis of the temporal lobe (RN) is a common late neurological complication due to proximity to the RT target, occurring in 4.
6-8.
5%.

There is evidence that RN is permanent and irreversible and may progress over time, which will cause a range of neurological symptoms such as memory loss, cognitive impairment, and dysfunction of attention and visual perception, which can seriously affect the quality of life
of patients.
Therefore, it is valuable to accurately predict RN in the acute or early delayed phase after RT (which is reversible less than 6 months after RT).

Recent reports point to early RT-associated brain alterations as early warning signs
of neurocognitive dysfunction in the late delayed phase after RT.
However, whether brain changes associated with RT in the early stages can be used to predict RN in the late delayed stage remains an open question
.
Therefore, understanding the longitudinal evolution of brain structural changes in NPC patients from acute to late delayed phase after RT and its relationship with RN is of great significance
for early identification and prevention of RN.

White matter (WM) in the brain has been reported to be more susceptible to radiation damage than gray matter (GM), possibly due
to the relatively scarce blood supply to white matter.
There is increasing evidence that radiation-related WM microstructural changes and WM volume changes are potential imaging indicators of
radiation brain injury.

At present, machine learning has been widely used in the field of medical imaging, which can extract meaningful information from image data through different algorithms to classify and predict
at the individual level.
Recently, a study published in the journal European Radiology delineated different patterns of longitudinal WM volume changes in patients with and without RNA in patients with nasopharyngeal carcinoma, and used machine learning methods to construct a predictive model
of early WM volume changes after RN and RT.

This study followed 74 patients with asymptomatic nasopharyngeal carcinoma receiving standard radiotherapy for 36 months
.
Structural magnetic resonance images were collected at multiple time points in the first year after radiotherapy.

Longitudinal structure images are processed
with FreeSurfer.
Linear mixed models were used to outline different trajectories
of WM temporal changes between patients with and without RNA.
Four machine learning methods were used to construct a predictive model
for early RN and temporal WM volume change.

Patients with nasopharyngeal carcinoma with different outcomes (RN vs.
NRN) had different patterns of atrophy in the superior temporal gyrus (STG) after radiotherapy.

Patients with RN atrophy faster
than patients with NRN.
The predictive model constructed with temporal WM volume change in the early post-radiotherapy stage performed well on RN.
The area under the curve (AUC) was 0.
879 and 0.
806
at 1-3 and 6 months after radiotherapy, respectively.
In addition, the prediction model constructed with an absolute time volume of 1-3 months after radiotherapy also showed good performance.
The AUC was 0.
842 and validated by another independent dataset (AUC = 0.
773).

The density distribution plot shows the overlap of necrotic lesions in the whole sample of 15 patients in the RN group
.
RN, radiation-induced necrosis of the temporal lobe

It is understood that this study is the first study
to explore the predictive value of RN in the early stage (1-3 months and 6 months after RT) of time-based WM volume change in patients with nasopharyngeal carcinoma.
Different STG atrophy models
were observed in NPC patients with or without RN.
Although patients in both groups showed time-dependent atrophy in STGs, patients in the RN group had more pronounced
atrophy compared to the NRN group.
In addition, WM volume changes (or absolute volumes) in the early stages can predict RN occurrence in the late delayed phase after RT
.
The results of this study clarify the early detection and prevention
of RN after RT in NPC patients.

Original source:

Xiaoshan Lin,Zhipeng Li,Shengli Chen,et al.
Divergent white matter changes in patients with nasopharyngeal carcinoma post-radiotherapy with different outcomes: a potential biomarker for prediction of radiation necrosis.
DOI:10.
1007/s00330-022-08907-z