It might seem odd that a special issue about Gianni Bonadonna would publish a review on radiation-induced neurocognitive dysfunction. Dr. Gianni Bonadonna is considered a pioneer in medical oncology history, mainly due to new and revolutionary treatment approaches proposed in breast cancer and Hodgkin lymphoma. He had an active role in the field of medical oncology, especially through lectures and textbooks. He shared his considerable insight of understanding cancer behavior and evaluating research advances over the years to prevent tumor recurrence and preserve patients from unnecessary toxicity. From a theoretical point of view, this complex vision is valid for all malignancies and can be indirectly shifted to every primary cancer, including head and neck cancer (HNC). During the last decades, multidisciplinary treatment for HNC has improved clinical outcomes but makes acute and late toxicity challenging. This article highlights the main central nervous structures that have a major impact on the development of neurocognitive dysfunction after radiotherapy for HNC. We briefly summarize the specific structure contouring instructions and the dose-volume histogram parameters. The aim is to raise awareness in clinicians in defining normal tissues to optimize radiotherapy regimens.
Tumori 2017; 103(4): 319 - 324
Article Type: REVIEW
AuthorsFrancesca De Felice, Pierre Blanchard
- • Accepted on 13/07/2017
- • Available online on 25/07/2017
- • Published in print on 31/07/2017
This article is available as full text PDF.
The incidence and mortality of head and neck cancers (HNC) in Italy are approximately 9,300 and 2,820 patients per year, placing this disease at the 11th rank in terms of incidence and 15th rank in terms of mortality (1). Due to substantial improvement in locoregional control and survival rates of HNC patients, it is essential to understand, quantify, and prevent significant long-term radiation effects (2). In particular, a significant increase in the incidence of neurocognitive dysfunction has been observed, potentially related to the dose bath delivered using modern radiotherapy techniques to nontarget area of brain (3, 4).
The central nervous system consists of various anatomic structures, including temporal lobes, basal frontal lobes, olfactory bulbs, brainstem, hippocampus, hypothalamus, pituitary, and cerebellum. The hippocampus, the temporal lobes, the olfactory bulb, and the brainstem are the main compartments related to radiation-induced cognitive impairment, due to their close vicinity to target structures (5, 6). Radiation-induced toxicities largely depend on the dose received by these normal structures. Exceeding the tolerance of these structures could lead to brainstem dysfunction or brain necrosis (7). Accurate structures delineation is necessary, with potentially the use of subregions (7). However, neurocognitive dysfunction correlates with pathologic alteration after radiation therapy (RT), notably in the inferomedial temporal lobes (8). When using altered fractionation regimens, the quantity of brain area that receives fraction size >2 Gy was related to worst neurocognitive outcomes (7). Thus the delineation of central nervous system substructures could permit the inverse planning system to reduce high dose to these “new” critical sites.
The aims of this review are (1) to establish the main structures related to cognitive function and to summarize contouring details; (2) to explore the impact on cognitive impairment following RT; and (3) to identify where further radiation constraints are required and to potentially provide them.
Literature search strategy
A comprehensive search of PubMed database was performed, using the following combinations of research criteria: “neurocognitive dysfunction,” “cognitive function,” “neurotoxicity,” “brain injury,” “memory,” “attention,” “head and neck cancer,” “radiotherapy,” “radiation,” “organ at risk,” “sparing.” Only English-literature articles were considered. Preclinical and clinical studies of contouring organs at risk (OARs) in HNC were analyzed, as well as consensus guidelines. Additional references were selected from relevant articles. An attempt was made to include all relevant studies and systematic reviews. Search strategy was performed up to June 2017.
Structures involved in cognitive function
Physiologically, temporal lobe, hippocampus, perirhinal cortex, and brainstem play a predominant role in neurocognitive function. We briefly describe their anatomical and functional characteristics in adults.
Hippocampal structure sits in the medial temporal lobe and comprises the cornu ammonis and the dentate gyrus. Because of its curved shape, it has been conventionally analogized to a seahorse. In adult subjects, the volume of hippocampus is approximately 3.5 cm3 (9). It is a paired structure. Hippocampus has a crucial role in the consolidation of information from short-term to long-term memory and spatial navigation.
Perirhinal cortex is located between the medial temporal lobe and the ventral visual pathway. Anatomically and functionally it is associated to the hippocampus. It is involved in odor detection, discrimination, and olfactory memory, as well as associative flexibility and recall (10).
The temporal lobe fills the middle cranial fossa. The cortex of the temporal lobe is delineated by several major sulci and gyri (11). The superior, the middle, and the inferior temporal gyri form the lateral surface of the lobe, whereas the lateral occipitotemporal gyrus, the fusiform gyrus, and the parahippocampal gyrus constitute the inferior surface. The lateral surface is indented by the superior and the inferior temporal sulci; the inferior surface by the occipitotemporal sulcus and the rhinal sulcus. The superior border of the temporal lobe is defined by the sylvian fissure, which separates the temporal lobe from the frontal lobe and the parietal lobe above. The posterior part of the temporal lobe merges with the parietal lobe and the occipital lobe, without definite cortical landmarks. Temporal cortex is responsible for auditory, olfactory, vestibular, and visual senses, perception of spoken and written language, long-term memory, and affective behavior.
The brainstem is mostly located in the posterior cranial fossa and structurally continues with the cerebrum, superiorly, and the spinal cord, inferiorly. The brainstem comprises the midbrain, pons, and medulla oblongata. It mainly consists of nuclei and fibers descending from cerebral cortex and regulates vital cardiac and respiratory functions.
Accurate delineation of brain anatomic area is an extremely delicate operation and it is essential in order to maximize therapeutic ratio and reduce risk of neurocognitive dysfunction. Contouring details are shown in
Contouring delineation of the main central nervous system structures (anatomical landmarks)
|Organ at risk||Boundaries|
|Temporal lobe||Cranial edge of the sylvian fissure||Base of middle cranial fossa||Temporal bone, sylvian fissure, greater wing of sphenoid||Temporal lobe, tentorium of cerebellum||Temporal bone||Cavernous sinus, sphenoid sinus, sella turcica, and sylvian fissure|
|Brainstem||Third ventricle||Inferior limit of the foramen magnum||Mammillary bodies, prepontine and premedullary cisterns||Quadrigeminal plate, cerebellum, cisterna magna|
|Hippocampus||Anterior edge of the temporal horn||Atrium of lateral ventricle||Lateral ventricle||Uncus, medial edge of the uncal recess, quadrigeminal cistern|
|Olfactory bulbs||Orbitofrontal cortex||Cribriform plate|
The olfactory bulb appears as an oval-shaped structure and lies on the inferior side of the brain immediately above the cribriform plate of the ethmoid bone (17).
Evidence showing a dose-toxicity relationship
Increasing evidence has shown that RT for HNC is variably related with neurocognitive impairment. The main manifestations of neurologic deficits are decline in memory, attention, manual dexterity, and visual and verbal recall. However,a clear dose-toxicity relationship is poorly documented. Dosimetry to normal brain tissues has been determined in several series (18, 19). Details are listed in
Major studies reporting dose-toxicity relationship in head and neck cancer patients
|Author||Study type||Patients (n)||RT technique (n)||Dose-volume data||Dose-effect relationship|
|C = chemotherapy; D = dose; HNC = head and neck cancer; IMRT = intensity-modulated radiotherapy; NA = not applicable; NPC = nasopharyngeal carcinoma; RT = radiation therapy; PNSC = paranasal sinus carcinoma.|
|Hsiao et al (18)||Prospective||NPC (30)||IMRT (30)||Temporal lobes: Dmean 34.4 Gy; V60 9%||Decrease in cognitive function p = 0.017 p = 0.039|
|Gan et al (25)||Retrospective||HNC (10)||IMRT alone (5), with C (5)||Temporal lobe: NA||Memory impairment|
|Meyers et al (19)||Retrospective||PNSC (19)||3D RT (19)||Brain tissue: Dmax >60 Gy||Memory recall impairment p<0.03|
|Lee et al (20)||Retrospective||NPC (16)||3D RT (16)||Brain tissue: NA||Cognitive and intellectual deficits|
|Lam et al (21)||Retrospective||NPC (60)||3D RT (60)||Temporal lobe: NA||Visual memory deterioration|
|Cheung et al (22)||Retrospective||NPC (50)||3D RT (50)||Temporal lobe: NA||Lesion volume correlates with severity of general cognitive impairment|
|Cheung et al (23)||Retrospective||NPC (53)||3D RT (53)||Temporal lobe: NA||Learning and memory impairment|
|Hua et al (24)||Retrospective||NPC (27)||3D RT (27)||Brain tissue: NA||Deficits in auditory attention/concentration, immediate and delayed verbal recall, immediate visual recall, recent memory, higher-order visuospatial abilities, bimanual dexterity|
Proposed dose constraints
Safe brain RT constraints are still not precisely known. The Quantitative Analyses of Normal Tissue Effects in the Clinic (QUANTEC) report for the brain included RT-induced neurocognitive injury as an endpoint (27). The evidence of cognitive deterioration in radiation fields used to treat HNC is weak, as well as the evidence that brain total dose of 72 Gy in 2 Gy fractions is linked to irreversible neurocognitive dysfunction in adults is minimal (<5%). The influence of confounding variables should be considered. For instance, the large fraction size and the volume of brain irradiated, as well as the concomitant chemotherapy delivery, should represent the main variables that could influence the neurocognitive decline in HNC patients. It was estimated that large single dose per fraction (4.2 Gy/fraction versus 2.5 Gy/fraction) is associated with a significantly higher risk of temporal lobe necrosis (19% versus 5%) in 10-year nasopharyngeal cancer survivors (28). McDonald et al (29) reported a 3-year estimated 12.4% risk of temporal lobe radionecrosis (5.7% symptomatic) from a cohort of 66 patients receiving a median radiation dose of 75.6 Gy for skull base chordoma, chondrosarcoma, adenoid cystic carcinoma, or sinonasal malignancies. Temporal lobe maximum dose and V dose levels, including V40 >16.5 cm3, V50 >9.6 cm3, V60 > 5.5 cm3, and V70 >1.7 cm3, were significantly associated with the development of radiation necrosis. Moreover, it could be important to determine if RT-related cognitive impairment observed in patients is greater than age-related cognitive decline. The QUANTEC brainstem report proposed constraints for the brainstem (30). In total, 4 studies (2 of which included HNC patients) were selected to extract brainstem tolerance data to limit neurologic toxicity. Maximum brainstem dose of 50-54 Gy, V65 <3 mL, and V60 <5 mL were associated with low risk of neurologic deficit (31, 32).
Robust evidence for hippocampus and olfactory bulbs dose constraints is still lacking. The constraints proposed are based on single institute experience. Hippocampus maximum dose should be less than 10 Gy, and it is suggested to maintain V7.3 <40% of bilateral hippocampi volume to reduce radiation-induced cognitive impairment (15, 33).The maximum dose tolerable for the olfactory bulbs is 40 Gy (5).
Ideally, central nervous system structures should receive a dose exposure as low as possible without compromising target volume coverage. In order to reduce dose to nontarget brain tissue, new radiation technique approaches have been proposed. Due to its physical properties, proton therapy creates highly brain-sparing treatment plans. The potential benefits of intensity-modulated proton beam therapy in HNC are currently being investigated and high-level evidence is lacking. However, a recent dosimetric study has demonstrated a consistent lower mean dose to the central nervous system structures compared with IMRT treatment plan (34).
Describing dose-volume constraints could represent a key point in order to achieve a plan optimization, without compromising target coverage.
Dose constraints of the main central nervous system structures
|Organ at risk||Dose constraint|
|Top priority||Second priority|
|Dmax = maximum dose; V
|Temporal lobe||Dmax 72 Gy||V70 <1.7 mL|
|V60 <5.5 mL|
|V50 <9.6 mL|
|V40 <16.5 mL|
|Brainstem||Dmax ≤54 Gy||V65 <3 mL|
|V60 <5 mL|
|Hippocampus||Dmax <10 Gy||V7.3 <40%|
|Olfactory bulbs||Dmax = 40 Gy|
The majority of available data on radiation-induced neurocognitive dysfunction is based on retrospective series and therefore definitive conclusions cannot be drawn, due to confounding clinical factors and different analysis endpoints (3, 8, 18, 36). For instance, it remains difficult to discriminate whether the neurocognitive dysfunction is associated with high radiation dose to temporal lobe or hippocampus or both of them or to other structures. Recently, Ma et al (37) recognized late delayed injury in patients with nasopharyngeal carcinoma treated with IMRT. They analyzed functional connectivity alterations using functional MRI. They proposed that functional connectivity alterations between the cerebellum and cingulo-opercular network may be involved in the cognitive functional abnormalities. This indicates that radiation-induced neurocognitive dysfunction is not restricted to the previous mentioned area, but another encephalic region, such as cerebellum and cingulo-opercular areas, should be considered in OARs delineation, as implicated in RT-related fatigue (38). However, at present, for the sake of clarity, temporal lobe and hippocampus should be considered, and thus contoured as independent structures. Further clinical trials of central nervous system structures spared during HNC treatment are needed, in order to provide the real clinical benefit of their accurate delineation and to develop and verify the consistent value of dose-volume data. In HNC patients, to prevent neurocognitive dysfunction, the whole brain should be considered a heterogeneous structure for dose-volume calculations. The addition of brain substructures and thus of specific subregion dose constrains should become routine clinical practice, especially in nasopharynx and sinuses cancer, due to their anatomic relationship with central nervous system structures. This review highlights a major unanswered question about the dose constraints definition: the absence of dose constraint to substructures. Resolution of this issue should be important to better assess morbidity associated with HNC treatment. Cognitive decline can be exacerbated by comorbidities such as alcohol consumption, vitamin deficiencies, as well as age, chemotherapy regimen, and RT schemes, suggesting that these deficits might be treatment-related but could also be related to other patient-related factors (37). Publication of detailed brain structure dose-volume parameters, as well as clinical toxicity data, should be encouraged, in order to make data pooling possible. Primary HNC sites, as well as the heterogeneity of therapies employed, results in potential clinical bias. Investigators should collect more detailed neurocognitive evaluations in follow-up programs. New RT technique and new parameters for patient quality of life should be paramount in the treatment plan evaluation to reduce and prevent neurocognitive dysfunction.
- De Felice, Francesca [PubMed] [Google Scholar] 1, * Corresponding Author (firstname.lastname@example.org)
- Blanchard, Pierre [PubMed] [Google Scholar] 2, 3
Department of Radiotherapy, Policlinico Umberto I “Sapienza” University of Rome, Rome - Italy
Department of Radiation Oncology, Gustave Roussy Cancer Campus, Villejuif - France
Université Paris Sud, Université Paris-Saclay, Le Kremlin-Bicêtre - France