A Review of the Role of Neuroimaging in Neurotoxicity Monitoring in Children with Acute Lymphoblastic Leukemia

Introduction

Acute lymphoblastic leukemia (ALL) represents 25.40% of all new childhood cancer cases and is most frequently diagnosed in 1- to 4-year-old children [1]. In many developed countries, the 5-year survival rate exceeds 90% [2]. The prognosis for ALL is thought to be influenced by age at diagnosis (unfavorable ages being less than 1 year old or older than 10 years), sex (a higher risk in males), and white blood cell count at diagnosis (above 50 000 cells/μL being unfavorable) [3]. ALL treatment is lengthy, and approximately 40% of patients with ALL have ≥1 experience with treatment-related toxicity. Neurotoxicity occurs with a frequency of 3–28% [4]. Neurological complications change throughout the first 2 months of therapy, ie the induction phase [5,6]. Seizures, neurological deficits, and stroke-like symptoms are the most common neurological complications. Other complications include neuroinfections, posterior reversible encephalopathy syndrome (PRES), epilepsy, intracranial bleeding, systemic infection with neurological complications, hydrocephalus, aseptic meningitis, and methotrexate encephalopathy. Leukoencephalopathy can also occur as a complication after chemotherapy in ALL, with an incidence ranging between 17 and 85% [7]. The term is associated with damage to the white matter of the brain and refers to non-specific neurological symptoms (such as altered mental status) [8,9].

Prophylactic central nervous system (CNS) therapy in ALL includes intrathecal administration of methotrexate, high-dose chemotherapy, radiation therapy, or a combination of the above. It has been suggested that CNS complications of ALL may be of 2 types: CNS infiltration by leukemic cells, or secondary complications of the disease and/or its treatment. Among the chemotherapeutics used, the most significant neurotoxic potential is observed with methotrexate, vincristine, and L-asparaginase. Methotrexate acts as an antifolate antimetabolite [10]. As a result, the synthesis of RNA and DNA nucleotides is disrupted. In contrast, vincristine (a chemotherapeutic agent from the alkaloids of Catharanthus roseus) binds to tubulin, inhibiting cell division, and also interferes with RNA and DNA synthesis by blocking the use of glutamic acid [11]. Meanwhile, L-asparaginase reduces asparagine levels and deaminates asparagine. Leukemia cells are thought to be unable to synthesize asparagine de novo, and must acquire it exogenously [12].

In the peripheral nervous system, neurotoxicity manifests in chemotherapy-induced peripheral neuropathy [13]. In the CNS, however, neurotoxicity ranges from cognitive and intellectual impairment to encephalopathy and coma. Furthermore, children who survive ALL and undergo a stem cell transplant are at high risk for long-term peripheral neuropathy, with a predominant pattern of sensitization to fine fibers and pain [14]. Significantly, in addition to neurotoxicity among chemotherapy survivors, long-term neurocognitive impairment has also been observed [15], including cognitive impairment manifesting especially in attention and executive functions.

Addressing the role and potential benefits of longitudinal neurological screening is considered to be extremely important for those patients, given the prevalence of neurotoxicity. This may pose a serious challenge to public health [16]

It is worth noting that approximately 50% of children with ALL experienced low health-related quality of life, which did not improve after the onset of neurotoxicity [17]. Furthermore, the occurrence of neurotoxicity may be related to a poorer prognosis in pediatric patients with ALL [18].

The neurotoxicity of chemotherapeutic drugs is dependent on their ability to penetrate the blood-brain barrier [19]. The mechanism of their toxic effects is also based on inducing inflammation, oxidative stress, pro-apoptotic effects, inhibiting neurogenesis, and accelerating the aging of neurons. These effects are particularly pronounced in pediatric patients. This is because brain development occurs during childhood. Therefore, finding and testing methods to detect neurotoxic changes early is needed.

In addition, it has been shown that neurotoxicity as a complication after chemotherapy in children with ALL can be associated with various pharmacogenetic variants. For example, the presence of the rs2306283 allele (of SLCO1B1) increases its risk threefold [20].

Another reason for neurotoxicity during ALL treatment may be increased oxidative stress [21]. Oxidative stress levels in the CNS are dependent on the intensity of chemotherapy. It is worth noting that the absence of cranial radiotherapy (CRT) as a component of overall therapy reduces the occurrence of changes in brain structure [19]. However, studies have shown that in childhood ALL survivors, these changes are also present after chemotherapy alone [22–24].

The latter consideration is particularly important because ALL can also affect the CNS, which occurs in approximately 30% of relapses and can impact the therapeutic success of treatment. No diagnostic methods are routinely used to check for CNS involvement or the initial neurotoxic effects of chemotherapy. Usually, tests are only performed when symptoms appear [22]. It is also suggested that radiological examinations in asymptomatic children with CNS leukemia are not clinically relevant at the time of diagnosis, but may be useful in patients with cranial nerve symptoms and negative cerebrospinal fluid results. In addition, they are considered to be important in monitoring those patients [23]. Early detection of neurotoxicity in one study led to improved neurological assessment documentation in pediatric patients [24].

Imaging studies such as CT or MRI can be instrumental in monitoring pediatric patients [25]. Their undoubted advantage is their non-invasive nature. CT scans are less detailed than MRI scans, but CT is cheaper and faster to use for diagnostic imaging. However, MRI is more commonly used for brain imaging. It is a technique used to study changes in the brain and its anatomy and to assess their effects. MRI images are taken without ionizing radiation, so patients are not exposed to its harmful effects [25]. Noteworthy is the lack of pediatric studies on the prognostic significance of radiological findings [26].

This article aims to review the role of neuroimaging in neurotoxicity monitoring in children with ALL. The study focuses on the role of imaging studies (MRI and CT) in selected conditions related to neurotoxicity from chemotherapeutics (methotrexate, vincristine, asparaginase), such as PRES, cerebral sinus venous thrombosis (CSVT), parenchymal/intracranial hemorrhage (ICH), and neuroinfections [26].

Imaging Studies in ALL-Related Neurotoxicity

One study evaluated the effectiveness of MRI in visualizing CNS lesions in ALL [27]. Based on the MRI images obtained, the patients were diagnosed with conditions such as PRES, CSVT, ICH, white matter lesions/leukoencephalopathy, CNS infection, sinusitis, middle ear infection, and brain atrophy. The toxicity of selected drugs and neurological disorders is shown in Figure 1. Thus, it has been shown that MRI can be functional for the early diagnosis of CNS lesions during ALL. CNS damage is one of the most common complications of ALL [28]. One study described brain MRI changes in children who completed therapy according to a protocol that included intrathecal chemotherapy or cranial irradiation as a method of ALL CNS prophylaxis. There were no significant differences in the patterns of white matter lesions between patients receiving either therapy. However, it occurred in half of the treated patients, was localized in the frontal and temporal lobes, and was diffuse. Moreover, there was no statistically significant correlation between white matter lesions visible on MRI and neurocognitive impairment in those patients [28].

By contrast, another study assessed cortical thickness and subcortical volumes from MRI and their correlations with behavioral measures in pediatric oncology survivors treated with chemotherapy [29]. They were observed to have thinner cortex in the right parahippocampal gyrus and smaller grey-matter volumes in the left globus pallidum, bilateral thalami, left caudate, and left nucleus accumbens, compared with controls. Those changes were shown to correlate with cognitive deficits. One more MRI study showed several changes during or after ALL treatment [30]. In addition, researchers suggest that this may be related to the mode of treatment. The following changes in the CNS during treatment or up to 3 months after ALL treatment were observed from MRI images: orbital, temporal, cerebellopontine angle, and spinal chloroma; bilateral subdural hematoma in the subacute stage; multifocal intraparenchymal hemorrhage; bilateral retinal hemorrhage and detachment; hematoma in the pons and mesencephalon; PRES bilateral leukemic infiltration of the 3rd, left 7th, and 8th cranial nerves; and meningeal leukemia. In contrast, after 3 months of treatment, MRI showed radiation necrosis and secondary brain tumor, osteomyelitis of the L3 vertebra, and meningeal leukemia [30]. In another study with MRI, the most common CNS-related complication of ALL was vascular disorders (58%), of which 34% were CSVT, 20% were PRES, and 4% of the patients had parenchymal hemorrhage [31]. This study suggested a diagnostic algorithm for detecting CNS lesions during ALL, which included conventional, post-contrast MRI, MRI venography, and diffusion-weighted MRI (DWI). CRT seems to be the most significant factor as regards the risk of CNS lesions [32]. In an MRI study of 25 patients, leukoencephalopathy was observed in 2 patients (one had grade III and the other grade I) (8%), cerebral atrophy in 2 patients (8%), old infarction in 1 patient (4%), and old hemorrhage in 1 patient (4%). Similar observations from MRI in children undergoing CRT were made in still another study [33]. It is noteworthy that survivors of CRT showed reduced white matter microstructural integrity (reduced fractional anisotropy) on MRI and associated neuropsychological dysfunction 25 years after treatment [34]. In addition, those individuals showed accelerated brain aging and an increased risk of early dementia.

Numerous case reports of complications during ALL treatment of children with methotrexate and vincristine can be found in the literature (Tables 1, 2) [35–48]. Zhang et al reported that a head CT scan for a patient showed no lesions, while an MRI scan could not be performed due to the patient’s lack of cooperation [49]. In Table 1, symptoms of neurotoxicity have been grouped into categories: altered mental status/encephalopathy (sleepiness, disorders of consciousness, confusion, disorientation, encephalopathy, disinhibition, agitation), seizures (focal, generalized, convulsions, faciobrachial seizures), systemic symptoms (fever, pallor, general weakness, vomiting, respiratory failure), meningeal signs/neuroinflammation (back pain, arachnoiditis, bulging anterior fontanelle), focal neurological deficits (hemiparesis, monoplegia, facial paralysis, flaccid paralysis, limb weakness, polyneuropathy, hypoesthesia, numbness, paresis, gait disturbances, stuttering symptoms, transient weakness), language and motor speech disorders (aphasia, speech disorders, oro-lingual apraxia), sensory disturbances/paresthesia (paresthesia of fingers/tongue, numbness), sensory hallucinations/visual disturbances (visual loss, auditory hallucinations, visual hallucinations), and ICH [35–47].

The above studies emphasize that MRI provides high sensitivity in detecting complications associated with neurotoxicity in the course and treatment of ALL [27–48]. Furthermore, DWI, MRI venography, and morphometric analyses have revealed micro/structural changes associated with neurocognitive deficits.

PRES

PRES is a neurological disorder in which epileptic seizures, headaches, impaired consciousness, and visual disturbances are observed. It can manifest itself in an acute or subacute manner, with symptoms developing over a period of several hours to several days. It is worth noting that the syndrome is usually reversible. However, it can also lead to various complications, which may be associated with cognitive disorders or focal neurological deficits, as well as cerebellar herniation through the intervertebral foramen. This syndrome is usually diagnosed based on characteristic features visible in MRI images [50]. MRI shows hyperintense lesions in T2 and fluid-attenuated inversion recovery (FLAIR) sequences, images in DWI or apparent diffusion coefficient (ADC) sequences show diffusion restriction within those lesions [51,52]. In addition, imaging studies (CT, MRI) usually reveal extensive brain edema, which often predominates in the parietal and occipital regions, followed by the frontal region and the temporal-occipital junction and cerebellum. If the edema increases, there is a risk that the foci of the disease merge [53]. Common factors that can cause this syndrome include: blood pressure fluctuations, preeclampsia/eclampsia, renal failure, and autoimmune diseases. Yet, it is also observed in the case of administration of cytostatic drugs [54].

In a study by Thavamani et al, PRES occurred in 0.04% (n=825) of children hospitalized in the United States in 2016 [55]. The single-center study, which included 24 children, showed that chemotherapy resulted in PRES in 66.7% of the patients, the average age at diagnosis being 6 years [56]. In a study by Bilir et al, PRES was mostly diagnosed in patients with malignant disorders [57]. The mean age of the patients was 8.16±3.77 and 8.95±3.66 years. A total of 38 children (65.52%) were diagnosed with ALL; 29 children with B-cell ALL, and 9 with T-cell ALL. Over half of the patients were in the treatment induction phase. The clinical symptoms were hypertension, focal and generalized convulsive seizures, altered consciousness, visual impairment, speech disorders, and hearing difficulty. On MRI examination, the following were observed: hyperintense lesions in the T2A series, with the relevant areas being the occipital, parietal, and frontal lobes (45%). On follow-up MRIs, in 40 of 58 patients, normal results were obtained in 20 patients, regression was observed in 18 patients, and progression in 2 patients [57].

A study by Banerjee et al included 649 children treated with Nordic Society of Paediatric Haematology and Oncology (NOPHO) ALL-92 and ALL-2000 high-dose methotrexate [5]. As many as 4.46% of the patients developed PRES symptoms, and 1.54% showed hypertensive encephalopathy. A comparison of the incidence of CNS symptoms with the occurrence of epileptic seizures is presented. All patients with PRES and hypertensive encephalopathy had epileptic seizures on imaging [58]. According to the researchers, as many as 73% of the patients with CNS symptoms had epileptic seizure episodes. PRES and hypertensive encephalopathy are overlapping syndromes so, in many cases, patients with PRES also meet the criteria for hypertensive encephalopathy. In the present study, patients with PRES and hypertension met the criteria for hypertensive encephalopathy, whereas not all patients with hypertensive encephalopathy met the criteria for PRES, because they either did not have an imaging study performed, or only had a CT scan done without typical radiological findings [5]. The diagnosis of PRES is most often made in the first month after the onset of symptoms, while the diagnosis of encephalopathy is usually formed in the following months [5]. In the work of Baytan et al, of 323 children with ALL, 6.50% had signs of neurotoxicity, of which 28.57% had PRES and 14.29% developed methotrexate encephalopathy [58]. Patients with PRES showed extensive bilateral hyperintensity involving the parietal-occipital region and other cortical and subcortical areas in T2-weighted images with inversion posterior reversible leukoencephalopathy. Axial T2-weighted imaging with inversion reconstruction after FLAIR showed T2 signal abnormalities in the bilateral white matter and grey matter of the parietal-occipital region and right frontal region. DWI showed diffusion restriction in areas of FLAIR abnormalities. Patients with encephalopathy (14.29%) showed bilateral diffusion restriction in the white matter. Brain abscess occurred in 5 children with neutropenia [58].

In general, it is believed that CT may not be sensitive enough for the early diagnosis of PRES. Patients with negative CT results but showing symptoms of PRES should undergo MRI to confirm or rule out the diagnosis [59]. During the course of ALL treatment in children, various changes in areas of the brain are observed, including PRES. Due to the fact that symptoms of the syndrome may be reversible, a rapid diagnosis is necessary. Late diagnosis worsens the patient’s prognosis in terms of reversibility of the changes.

CSVT

CSVT is a rare neurological condition that can lead to a high mortality rate [60]. The disease is characterized by a variety of symptoms and risk factors, which makes initial diagnosis very difficult. In terms of clinical manifestations, there are 2 categories: those associated with increased intracranial pressure attributed to impaired venous drainage, and those associated with focal brain damage due to venous ischemia/infarction or hemorrhage [61], depending on the location of the clot. The usual symptoms are headaches (diffuse and increasing), double vision, seizures (in about 40%), or other neurological symptoms.

Neuroimaging is the gold standard for the diagnosis of CSVT [62]. CT without contrast is the recommended initial imaging protocol. However, it is worth mentioning that approximately 30% of patients may not have any visible lesions. The most preferred methods are MRI with contrast and magnetic resonance venogram (MRV). However, the MRI technique also has its limitations. One study noted that MRI showed signs of CSVT in only 1.5% of all cases [63].

Imaging is usually performed approximately 7 days after clinical symptoms appear [64]. Direct symptoms include “positive” symptoms, ie, visualization of the clot on MRI and CT, and “negative” symptoms, ie, visualization of thrombotic material in the form of filling defects on CT venography and MRV. Intermediate signs may include changes within the brain parenchyma, eg, venous edema, venous infarction, subarachnoid hemorrhage, and parenchymal hemorrhage.

Yet, it is worth noting that MRI results will vary according to the age of the thrombus, as signal characteristics change during the acute and subacute stages, depending on the presence of blood breakdown products [65]. Among other things, thrombosis can occur after the use of various chemotherapeutics. For example, asparaginase has received a lot of attention in the literature [66]. It is thought to cause a deficiency of natural anticoagulants, which promotes thrombus formation.

Also, steroids administered during chemotherapy to patients with ALL may increase the risk of thrombosis [67]. For example, administration of prednisolone in healthy individuals induced a procoagulant state through increased thrombin generation and increased levels of plasminogen activator inhibitor type 1 (PAI-1) and von Willebrand factor. Another study revealed that the risk of thrombosis resulted from the combined effect of glucocorticosteroid use and the underlying disease [68].

Still, several studies have repeatedly demonstrated the effect of asparaginase and identified other risk factors for CSVT. Significantly, of the 91 thrombotic events of patients with ALL, 26 were associated with CSVT [69]. It was also shown that patients taking L-asparaginase at lower doses and for longer periods of time showed a higher risk of thrombotic events.

Moreover, the median time between L-asparaginase treatment and onset of symptoms was shown to be 21 days (incidence rate 1.5%) [70], while the mortality rate was 28%. In a cohort study of a population of children (aged 6–14 years) from the United Kingdom diagnosed with ALL, CSVT was found to occur at a similar frequency, 1.4% [71]. Importantly, it was only observed with pegylated (PEG) asparaginase. Furthermore, it was found that older age was associated with an increased risk of CSVT.

Another study of children treated according to the NOPHO-ALL 2008 protocol showed that the frequency of CSVT complications was 2% [72]. In addition, 16 of those were associated with the use of PEG-asparaginase and the remaining 16 with the use of steroids.

Furthermore, still another study using the NOPHO treatment protocol brought about similar results. The difference was that the study population included individuals aged 1–45 years [73]. However, it is noteworthy that the cumulative incidence of CSVT was highest among adolescents, at 1.9%.

In contrast, one study estimated the incidence of CSVT in ALL to be 6.2% [74]. The majority of CSVT cases (77%) occurred during the post-induction phase of treatment with a combination of asparaginase and dexamethasone. An association between older age, T-cell immunophenotype, and a higher disease risk was also confirmed. A study of a Chinese population (n=7640) showed that out of 28 patients with cerebral thrombosis, 20 developed CSVT [75]. The involved sites included the upper sagittal sinus, transverse sinus, sigmoid sinus, straight sinus, and torcular herophili. In this study population, too, the patients mainly received asparaginase.

It is noteworthy that the form of asparaginase is also relevant to CSVT risk [76]. A retrospective study considering children diagnosed with ALL and lymphoblastic lymphoma showed that CSVT only occurred in patients who received native L-asparaginase or PEG-asparaginase. Erwinia-derived asparaginase did not cause CSVT in those patients. However, a high CSVT incidence (10.5%) was observed with the L-asparaginase and PEG-asparaginase forms. Triglycerides >615 mg/dL, mediastinal mass, and larger body surface area were also shown to be risk factors associated with the occurrence of CSVT. Among the drugs that the patients received, prednisone, vincristine, daunorubicin, asparaginase, cyclophosphamide, mercaptopurine, and cytarabine can be distinguished in terms of their potential risk [76].

An interesting study assessed the effect of CSVT on neurocognitive processes [77]. The study group consisted of children diagnosed with hematological malignancies and CSVT (all subjects had received L-asparaginase in combination with other drugs). Cognitive and behavioral tests were performed. The study showed that the subjects achieved good, average, or above-average scores on those tests. A few had problems with non-verbal memory, visuomotor integration, and anxiety [77].

In one case report, a boy diagnosed with ALL was treated with prednisolone and daunorubicin, and received weekly intravenous infusions of vincristine and intrathecal chemotherapy [78]. In addition, on the third day of the induction protocol, he received PEG-asparaginase and E. coli polyethylene glycol L-asparaginase conjugate (2500 IU). The first symptoms occurred after 24 days. A CT scan showed a subcortical hemorrhage. In contrast, MRI showed bihemispheric multifocal hemorrhagic infarcts with vasogenic edema. Post-contrast images revealed a filling defect in the superior sagittal sinus. MRV confirmed the changes observed on MRI. A 4-year-old boy with ALL received methotrexate, vincristine, dexamethasone, epirubicin, and L-asparaginase, among others [79]. After an intramuscular injection of the latter, the boy initially developed a rash and then experienced neurological symptoms in the form of seizures and hemiparesis. A CT scan showed an essentially normal brain. MRI was also performed and showed foci of filling defects in the lumen of the superior sagittal sinus with an “empty delta” sign on T2-weighted images. Thrombosis of the superior sagittal sinus was found. Also, in another case report, after chemotherapy, a 9-year-old boy developed CSVT before and after ALL treatment [80]. The first MRI scan of the patient’s head and neck vascular plaque showed thrombosis of the right internal jugular vein, transverse sinus, and sigmoid sinus. A second examination after the diagnosis of ALL, using MRV, confirmed the MRI result and, after treatment, the patient was additionally found to have thrombosis of the right sigmoid sinus, transverse sinus, and sagittal sinus. The patient had received vincristine, daunorubicin, L-asparaginase, and corticosteroids as treatment [80].

In yet another case report, a 3-year-old girl developed CSVT before the diagnosis of ALL [81]. The patient was brought to the emergency department 3 weeks after discharge from hospital. She was experiencing hallucinations and did not recognize her parents. MRI and MRV, among others, were performed, and she was found to have thrombosis in both the left sigmoid and the transverse sinus. She received treatment in the form of low-molecular-weight heparin and 2 weeks later was admitted to the hospital again. At that time, ALL was diagnosed, and she was given L-asparaginase. Imaging examinations performed during treatment did not identify recurrent CSVT [81].

Interestingly, one case report describes CSVT in a 12-year-old boy with ALL after intrathecal administration of methotrexate [82]. It is noteworthy that the boy had an indeterminate and homozygous C677T mutation of methylenetetrahydrofolate reductase, which may have contributed to this particular disease. Moreover, in that case, no changes were noted on CT. An MRI scan was performed in which the following were observed: restricted diffusion in the bilateral centrum semiovale (deep cerebral white matter) with confluent non-enhancing areas with T2/FLAIR hyperintensities and absence of normal flow void in the left transverse and sigmoid sinuses [82].

The drug that most commonly caused CSVT was asparaginase, but the condition also occurred with other treatments. Furthermore, the occurrence of CSVT was not always related to chemotherapy itself but also occurred before the diagnosis of ALL [83]

Parenchymal/Intracranial Hemorrhage

Extravasation of blood both in brain parenchyma and in the surrounding areas located in the skull cavity can be classified as an ICH [84]. Symptoms usually depend on the size of the hemorrhage and intensify within a few minutes to a few hours.

Typical symptoms include headache, vomiting, and nausea. In cases of large hemorrhages, decreased consciousness, coma, and seizures are also reported [84]. A deterioration in the patient’s neurological condition may indicate an enlargement of the hematoma or worsening of the edema. Depending on the location of the hematoma, symptoms may also include sensory-motor deficits, aphasia, vision problems, and cognitive dysfunction [85]. ICH is a phenomenon which may result, both in the acute and subacute phase, in permanent neurological deficits, such as complications of hemiplegic paralysis or cognitive function disorders, which is why early diagnosis and prevention are so important [86].

In a boy with ALL treated with methotrexate, who had 2 notable genetic polymorphisms: Ikaros family zinc finger 1 (IKZF1) deletion and positive Breakpoint Cluster Region-C-Abelson oncogene 1 (BCR-ABL1), ICH was observed during maintenance treatment. This resulted in neurological consequences, including hemiplegia [87].

CT is one of the methods for diagnosing intracranial bleeding, especially in acute conditions with the presence of macroscopic hematomas [88]. Its advantages are considerable, including availability and short duration of study. MRI is superior to CT, however, in terms of detection of microhemorrhages, as well as the possibility of a more precise characterization of the hemorrhage, including the time of its occurrence [88].

MRI is considered to have approximately 100% sensitivity and specificity, as well as accuracy in diagnosing ICH. The results will depend on hemorrhage age, as hemoglobin will transform over time into oxyhemoglobin, deoxyhemoglobin, methemoglobin, and ultimately into ferritin and hemosiderin [88,89]. For example, in the acute and hyperacute phases in T1, the image of the hemorrhage is hypointense/isointense, and then in subsequent phases it is hyperintense (early/late subacute phase) [89]. In the T2 image, there is an isointense/hyperintense center with peripheral hypointensity and a hyperintense rim of vasogenic edema in the hyperacute phase, then hypointense with hyperintense rim in the acute phase, and hypointense in the early subacute phase, which transitions into hyperintense in the late subacute phase [89]. Factors increasing the risk of ICH include hypertension, cerebral amyloid angiopathy, male sex, older age, Asian ethnicity, high daily alcohol intake, low cholesterol levels, genetic factors, taking anticoagulants, and drug abuse [84,90]. However, it may also occur as a result of the medications used as part of anticancer therapy in patients with ALL, though it is not a common complication. Rahiman et al presented the results of their study covering patients from the last 10 years, which included over 923 patients treated in accordance with the United Kingdom 2003 guidelines for ALL (UK-ALL 2003). Neurotoxicity occurred in 10% of the people treated for ALL, and intracranial bleeding occurred in only 5 people, representing 0.5% of the patients studied. This complication solely occurred during the induction phase of treatment. Only 1 patient had a fatal outcome but it was related to a systemic infection [6]. Ulu et al, however, reported numerous cases of neurotoxicity, including isolated cases of hemorrhagic complications, such as bilateral frontoparietal subdural hematomas in the subacute stage, hematoma in the pons and mesencephalon, and multiple focal intraparenchymal hemorrhages [30,91]. The use of methotrexate, vincristine, and asparaginase is suspected to be responsible for the occurrence of neurotoxicity in the examined patients, but determining exactly which drug caused which complication is difficult.

L-asparaginase is one of the drugs that have been proven to cause bleeding complications. However, it is much more likely to cause thrombosis in the CNS. The greatest sensitivity to this drug is seen in the consolidation and re-induction phases of treatment [92]. Most bleeding complications also occur in the CNS [93]. The pro-hemorrhagic effect of L-asparaginase results from the depletion of plasma proteins, including coagulation factors, disturbing coagulation and fibrinolysis processes [94]. Repeated doses of L-asparaginase deepen the deficits of coagulation factors, intensifying the bleeding problem. Data indicate that ways to counteract those effects are to use fresh frozen plasma, cryoprecipitate, platelets, anticoagulant therapy, or steroid drugs [93]. The risk of death is high, and surgical treatment is not superior to pharmacological treatment [94,95].

Methotrexate is another important drug with numerous adverse effects but still remains one of the basic drugs in ALL therapy [87,96]. There are reports that it can also cause intracranial hematoma, especially when administered intrathecally; however, Ning et al presented a case of an 11-year-old patient diagnosed with acute B-cell-progenitor ALL with the BCR-ABL1 fusion gene and IKZF1 deletion, which could have caused an increased risk of serious complications of therapy. This boy received induction treatment, followed by 2 cycles of intravenous high-dose methotrexate therapy. Severe adverse effects occurred, including myelosuppression, rash, and liver function disorders, but an intracranial hematoma only appeared after oral administration of a small dose of methotrexate during maintenance chemotherapy. The boy’s condition improved dramatically after discontinuing oral methotrexate and implementing supportive therapy. This demonstrates the need to closely monitor high-risk patients, to protect them from adverse outcomes [96].

The appearance of ICH in imaging is highly variable, reflecting the diversity of pathologies that cause ICH, but neuroimaging is one of the key diagnostic methods for this condition and for predicting its course. Having the speed of CT scans in mind, MRI should not be the only method used in the diagnosis of ICH [97].

Neuroinfections

Neurological disorders associated with neuroinfection pose a serious challenge for patients and society [77,78]. These disorders include dementia, motor neuron diseases, sleep disorders, and peripheral neuropathies. In the course of neuroinfection, there is increased inflammation and infiltration of immune cells into the CNS, as well as chronic activation of astrocytes and microglia. In addition, increased expression of chemokines and cytokines is observed [98]. According to the accepted definition, neurotoxicity means any adverse effect on the structure or function of the peripheral or CNS resulting from the action of various factors, including biological factors [99,100]. The most common clinical symptoms in children include fever, impaired consciousness, and convulsions. Vomiting, headaches, and meningeal symptoms may also occur [101].

Recognizing neuroinfection seems to be quite difficult because the infected lesions are difficult to sample. MRI is considered the primary imaging method. These lesions can be classified based on their characteristics in T1, T2, and contrast enhancement. They include ring-enhancing lesions, basal ganglia space-occupying lesions, gray-matter hyperintensity, and white matter hyperintensity [102].

However, it is worth noting that in one study in which CT, MRI, and transcranial ultrasound were performed, 70.5% of the patients showed no changes in imaging tests, yet developed a neuroinfection [101,103]. Opportunistic infections are frequently found in ALL pediatric patients, due to neutropenia and a weakened immune system, caused by both the disease itself and systemic treatment [104]. It is believed that most complications [105] occur during the induction period [58]. In another study, it was shown that infections were the main cause of death in cases of acute neurological complications [6].

Other adverse prognostic factors are reduced nutrient intake, prolonged hospitalization, indwelling catheters, and ulcerative lesions prompted by mucotoxic chemoradiotherapy, among others [106]. Common infectious agents responsible for neuroinfections include fungi (Aspergillus, Candida) [6,107–109] and viruses (Herpes simplex virus type 1 and 2) [106]. The pathogens may reach the brain parenchyma (focal or diffuse), the meningeal lining, or the extra-axial space, with disease being more dispersed and treatment-resistant. Focal parenchymal infections most commonly take the form of cerebritis which, when left unmedicated, develops into an abscess [85]. Neuroinfections account for a high mortality rate of up to 74% [108,109]. Baytan et al retrospectively analyzed hospital reports of 323 pediatric patients with ALL for acute neurological complications. Among 23 patients who presented with neurological impairment, fungal infections were the second most common cause (n=5/23; 21.7%). Most CNS complications occurred during the induction period. Abnormalities have been seen on MRI in 19 of 23 patients (82.6%) [58]. In another retrospective study, neuroinfections were identified in 18 (20%) out of 90 patients with neurological symptoms. Involvement of the nervous system included a brain abscess, invasive rhinocerebral mucormycosis, encephalitis, and meningitis. The most-often identified pathogens were Aspergillus and Herpes simplex virus; however, in a vast group of patients, the symptoms were of unknown etiology [58]. It is worth noting that infections are associated with some of the most common neurological complications during ALL [107,110]. For example in one of the studies, meningitis was the most common neurological complication in children with ALL [111]. Apart from retrospective multicenter analyses, the available literature discussing neuroinfections in pediatric patients with hematologic malignancies contains single-institution reports describing bacterial meningitis caused by Yersinia enterocolitica [112], Campylobacter fetus [113], or Stomatococcus mucilaginosus [114], as well as disseminated Cryptococcus fungemia [115], among others.

The imaging findings of neuroinfections in pediatric patients with acute leukemia do not differ from those in the general population [116]. The most commonly used compounds for contrast enhancement in suspected CNS infections are gadolinium-based contrast agents. MRI spectroscopy and DWI may assist in further differential diagnosis [107]. Even though most pathogens reach the CNS via the bloodstream, infection may occur directly from a nearby infection site, such as the sinonasal region, which should be examined throughout [117]. Pathologies in the brainstem can also be detected on MRI [118]. Aspergillus infection of the nervous system is frequently angioinvasive, resulting in parenchymal hemorrhages [119], and occurs in the form of abscesses, whereas Candida infection can cause ring-enhancing lesions [107]. Advanced parenchymal destruction correlates with high mortality [120]. Pyogenic abscesses show an isointense-to-slightly hyperintense rim on T1-weighted images and are hypointense on T2-weighted images [119], while fungal abscesses appear as solid or ring-enhancing lesions [107,121]. Most cases of viral CNS infections in the form of encephalitis are comparable, both clinically and diagnostically. Herpes encephalitis can reach the limbic system and brainstem [119].

When it comes to neuroinfections, it is key to figure out what is causing the infection, so the right treatment can be given. In patients with ALL, it is often opportunistic pathogens that can lead to life-threatening infections. Imaging is not the main method to figure out what is causing the infection, but it can help.

Limitations of our Study

Case studies are a valuable research tool. An unquestionable advantage of case reviews is their ability to present new clinical findings regarding diagnostic methods or symptom observation [122,123]. However, they are not without their drawbacks. The main ones are the fact that these studies may not be representative of the entire population of children with ALL treated with chemotherapy who are experiencing neurotoxicity. It is also impossible to establish a cause-and-effect relationship. Furthermore, misinterpretation of the available literature and bias in publications should be taken into account. It is worth noting that some case reports are written retrospectively, rather than during ongoing patient observation, and therefore may not reflect the full clinical picture of the case described. To reduce the limitations of a case review, this study was supplemented with a narrative review of the available literature, which expanded this work to include other studies and allowed for stronger conclusions to be drawn. It should be noted that the adult population was excluded from the present study, but this does not rule out the use of data on adult patients from previous studies to understand neurotoxicities that occur in children. It is worth noting that some studies did not consider simultaneous CT and MRI, making it difficult to assess their simultaneous utility.

Future Directions

The importance of testing and taking test results into account in the treatment of ALL and CNS complications is worth emphasizing, due to the likelihood of complications and their possible early prevention. Imaging tests are not usually included in ALL treatment protocols. However, it would be worthwhile to introduce them into clinical practice for the prevention of complications in children with ALL treated with methotrexate, vincristine, or L-asparaginase to confirm or rule out neurotoxic CNS complications more quickly. This would certainly reduce the severity of neurotoxicity-related complications in these patients and improve their quality of life. In the future, research should focus on the role and use of artificial intelligence algorithms that could help with image analysis, influence the decision-making process, and reduce the risk of diagnostic errors, thereby improving the prognosis for pediatric patients with ALL. Due to the limited number of studies, it would also be advisable to construct predictive models that use CT and MRI images to predict the risk of neurological complications, which would allow for early prevention in patients at risk.

Conclusions

Early signs of neurotoxicity after methotrexate, vincristine, and asparaginase therapy seen on MRI and CT imaging include leukoencephalopathy, encephalopathy, and PRES. Changes mainly include reduced restrictive diffusion and periventricular hyperintensity in the subcortical white matter, as a result of cytotoxic swelling caused by ALL treatment. Occasionally, lesions may involve the frontal, parietal-occipital, or occipital subcortical white matter. Neurotoxicity is particularly dangerous when treating children, due to different drug absorption and metabolism, compared with adults. By choice, MRI examination is more accurate in diagnosing neurological lesions; however, it is not always available, especially on-call and on an emergency basis. For this reason, we usually perform a head CT scan first, then follow up with an MRI scan. A head CT scan is more accessible, less costly, and performed faster than an MRI scan. Due to the excessive length of the MRI examination, pediatric patients often need to go under general anesthesia and their clinical condition does not always allow this, especially when patients have neurological symptoms. These imaging methods allow for non-invasive differentiation of the causes of neurotoxicity and implementation of appropriate treatment.