Comprehensive review of post-treatment imaging in head and neck cancers: from expected to unexpected and beyond

Abstract

Head and neck cancer management requires multidisciplinary approach in which radical surgery with or without flap reconstructions and neck dissection, along with radiotherapy (RT)/chemoradiotherapy (CRT) serve as the key components. Neoadjuvant chemotherapy and immunotherapy are used in selected cases based on the institutional preference. Knowledge of expected post-treatment changes on imaging is essential to differentiate it from recurrence. In addition, awareness of various post-treatment complications is imperative for their early detection on imaging. Distorted anatomy after treatment poses diagnostic challenge, hence, proper choice of imaging modality and appropriate timing of scan is pertinent for accurate post-treatment evaluation. In this article, we have comprehensively reviewed expected post-treatment appearances and complications on imaging. We have discussed imaging appearances of recurrences at the primary and lymphnodal sites and discussed documentation of findings using Neck Imaging Reporting and Data Systems (NI-RADS). We have also delved into the patterns of recurrence in human papillomavirus (HPV) positive HNSCC. Furthermore, we have provided flowcharts and discussed recommendations on the site-specific and treatment-related imaging modalities to be used along with their appropriate timing, for adequate evaluation of HNSCC after treatment. In addition, we have also touched upon the role of advanced imaging techniques for post-treatment HNSCC evaluation.

Introduction

Lip and oral cavity cancer ranks highest in cancer incidence amongst head and neck squamous cell carcinomas (HNSCC), with 2% new cases worldwide as per GLOBOCAN 2020 data, followed by laryngeal cancer with 0.96% new cases.¹ Tobacco and alcohol consumption are responsible for approximately 50%-60% of HNSCC.^2,3 A strong association of human papillomavirus (HPV) with oropharyngeal carcinoma (OPC), and Ebstein-Barr virus with nasopharyngeal cancers is known.^2,4 About 50%-60% of treated HNSCC patients show locoregional recurrence within 2 years and 20%-30% of them develop distant metastasis.⁵

HNSCC management requires multidisciplinary approach with surgery and radiotherapy (RT)/chemoradiotherapy (CRT) being key components and neoadjuvant chemotherapy (NACT) having a selected role based on the institutional preference.^2,6–10 Checkmate 141 phase III trial, Keynote-040 and Keynote-048 randomized phase III trials have established the role of immunotherapy for recurrent and metastatic HNSCC, and currently the role of neoadjuvant immunotherapy is being explored.¹¹ Altered anatomy after radical surgery with or without flap reconstruction and neck dissection, post-RT/CRT/NACT/immunotherapy expected changes, recurrence, and post-operative or post-RT complications, all pose diagnostic challenges to the radiologist evaluating post-treatment scan.

In this article, we have extensively reviewed post-surgery and post-RT expected changes and complications on imaging, along with post-RT/chemotherapy/immunotherapy response evaluation, recurrence and metastasis detection in HNSCC, and discussed post-treatment assessment using Hopkins criteria and Neck Imaging Reporting and Data Systems (NI-RADS). We have also discussed the current national comprehensive cancer network (NCCN) and European Society for Medical Oncology (ESMO) guidelines on post-treatment HNSCC imaging. Furthermore, we have formulated a flow chart on site-specific and treatment-related imaging modalities to be used along with their appropriate timing for adequate evaluation of HNSCC after treatment, and also provided a summary of imaging recommendations at the end, which can be quickly referred to by radiologists and clinicians to decide upon further line of management. In this article, we have also discussed patterns of recurrence in HPV positive HNSCC and enumerated role of advanced imaging techniques in post-treatment HNSCC evaluation.

Post-treatment imaging in HNSCC: expected changes and complications

Knowledge of expected/biologic changes on imaging after radical surgery with or without flap reconstructions and neck dissection, RT/CRT, and immunotherapy, is essential for differentiation from recurrence, and early detection of complications.

Post-surgical evaluation

Surgery is performed either for curative intent in early stage cancer (T1-T2 N0 [stage I and II] oral cavity, laryngeal, p16-negative oropharyngeal carcinoma or T1-T2N0 p16-positive oropharyngeal cancer as per the American Joint Committee on Cancer [AJCC]/International Union against Cancer [UICC] Tumour Node Metastasis [TNM] 8th edition), for local control in locally advanced HNC (T3/T4 oral cavity and T4 larynx carcinomas), for airway conservation, symptom amelioration, or in selected cases of advanced stage depending upon response to other first line therapies.^2,6,12 Advanced (T4) hypopharyngeal carcinomas with laryngeal cartilage invasion or non-functional larynx also warrant surgical treatment.² Salvage surgery, that is, surgery performed as a last resort after failed initial organ preserving treatment with concurrent CRT (CCRT)/RT, has overall high complication risk of approximately 67% and a locoregional recurrence (LRR) rate of 60% for oral cavity SCC, posing a significant challenge to the radiologists should the need for imaging arises after such a surgery.^13,14,Table 1 and Figures 1-5 show various types of HNSCC surgeries.^2,13,15–27 Contrast-enhanced CT (CECT) and contrast-enhanced MRI (CEMRI) are the predominant imaging modalities for post-treatment evaluation.

Imaging of post-operative complications

Post-operative complications occur early and may be categorized as follows: (a) Complications associated with radical surgery with or without flap reconstructions²²: This includes seroma; which appears as non-enhancing fluid collection, abscess; as a rim enhancing fluid collection showing diffusion restriction on MRI, haematoma; which appears hyperdense on CT and may show T1 hyperintensity and mixed signal on T2 due to blood on MRI, and pharyngocutaneous fistula (Figure 6); seen as a direct visualization of fistulous track communicating between the oral cavity/larynx and cutis. (b) Complications specific to flap reconstructions²²: This includes flap necrosis due to arterial or venous thrombosis seen as non-opacification of neck vessels on CECT with non-enhancement of flap and loss of flow void on T1 and T2W MRI with filling defect on CEMRI and doppler. Infection is another cause of flap necrosis seen as rim enhancing collections with air locules within the flap/surgical site. Implantable doppler and skin paddle monitor are used to monitor flap perfusion,^22,23 venous thrombosis is a common cause of flap failure and may be seen within 3 days of surgery.²² Radial forearm and fibular osteocutaneous free flaps are predominantly associated with distal limb ischaemia.²² Complications specific to neck dissection: This includes chylous fistula, which may occur in 1%-2% of the patients with level IV neck dissection, and is suspected when collections are located predominantly in left lower neck on imaging.²² Pre-operative RT, pre-operative CRT, malnutrition, anaemia, tobacco and alcohol use are risk factors for complications after head and neck surgeries.²²

Table S1 shows the various types of surgeries and the complications arising thereof.^{15,22,27–34}

Evaluation after RT/CRT

RT is a key component of HNSCC treatment, and clinical trials have shown that CCRT reduces risk of death by 19% and overall improved 5 year survival by approximately 8% in comparison to RT alone.³⁵ Definitive RT is the mainstay treatment for non-metastatic nasopharyngeal carcinoma.³⁶ Definitive CRT may be considered in locally advanced inoperable oral cavity carcinomas.² CCRT is the treatment of choice for locally advanced HNSCC, especially, nasopharyngeal, oropharyngeal, hypopharyngeal, and laryngeal carcinomas.^2,6 High-risk factors (mainly for oral cavity carcinomas) for post-operative LRR; pT3-4 on resected specimen according to UICC/AJCC TNM 8th edition, tumour ≤1 mm from the margin (positive margin), tumour between 1 and 5 mm from the margin (close margin), perineural infiltration, lymphovascular involvement, >1 lymph node infiltration, and the presence of extranodal extension, entail adjuvant RT/CCRT.² In general, post-surgery adjuvant RT/CTRT is given when microscopic or gross residual tumour is suspected.⁶

Current RT techniques including intensity modulated RT (IMRT), 3D conformal RT (3DCRT), helical tomotherapy, volumetric modulated arc therapy (VMAT), and proton beam therapy (PBT), spare the critical organs from the radiation field, decrease normal tissue damage, and at the same time achieve primary objective of local tumour control.⁶

Imaging of expected/biologic effects after RT/CRT

RT affects tumour microenvironment by causing hypoxia, fibrotic responses, and immune activation, which in turn affects post-RT response to treatment.³⁷ CECT and Fluorodeoxy glucose Positron Emission Tomography CECT (FDG-PET/CECT) are the predominant imaging modalities for post-RT/CRT evaluation except for sinonasal and skull base tumours where MRI is the imaging modality of choice.⁶ CEMRI is also reserved for assessment of suspected perineural tumour recurrence, particularly in nasopharyngeal and skull base tumours.²²

Expected or biological changes after RT can be divided into early inflammatory and proliferative phases, and delayed tissue remodelling phase.³⁷ There is temporal evolution of post-RT biological changes in tissues within the radiation portal and Table 2 describes the early and late imaging findings (CT/MRI) along with underlying pathological changes that occur after RT/CRT.^21,22,38,Figure 7 shows expected early post-RT change on CT. On FDG-PET/CECT, early post-RT changes manifest as diffuse, symmetrical, low grade FDG uptake confined to the radiation field.³⁹ Groups of muscles within the radiation field may also show increased FDG uptake.³⁹

Imaging of post-RT/CRT complications

With the newer RT techniques (IMRT, 3DCRT VMAT, PBT), incidence of post-RT complications have reduced, though not nullified.^6,22 Technique and dose of RT, along with the volume of irradiated tissues, tumour location and stage, and poor oral hygiene have bearing on post-RT complications, and alcohol and smoking are known to aggravate RT related complications.²² Early post-RT complications like oral mucositis and skin desquamation may get exacerbated due to CCRT and NACT.^22,40,41 Post-operative complications occur more after CRT (46%-100%) than after RT alone (37%-74%).^22,42 The various post-RT complications along with their pathologic basis are enumerated in Table 3.^{21,22,37,38,43–50} and their imaging features are described in Table 4.^{21,22,38,42,43,51–53} CT images of osteoradionecrosis (ORN) is shown in Figure 8. It should be noted that osteoradionecrosis may show diffuse inflammation with intense FDG-PET/CECT uptake in the tissues surrounding the lytic and sclerotic bone (Figure S1), however, a focal enhancing soft tissue is absent.

Flowchart in Figure 9 enumerates the imaging recommendations for suspected RT related complications.

Induction chemotherapy/NACT

The main objectives of NACT in locally advanced HNSCC are organ preservation (laryngeal or hypopharyngeal squamous cell carcinomas), tumour shrinkage to render it resectable, reduce surgical margins, reduce distant metastasis, and reduce intensity of subsequent RT/CRT.^2,7–9,54 Febrile neutropenia, diarrhoea, and mucositis are the common complications after induction chemotherapy, which are diagnosed clinically and do not require imaging.⁷

Response assessment after NACT on imaging

Responders to NACT tend to respond better to subsequent RT and have an improved overall survival (OS), hence, adequate response assessment on imaging is of paramount importance.⁸ On post-NACT imaging, if the tumour shrinks from sites which are critical for determining resectability, then tumour is considered to be operable and surgery is performed (with adjuvant CRT/RT).^9,10 Conversely, if the tumour persistently or newly involves a critical structure after NACT on imaging, the tumour is inoperable. Those who are inoperable after NACT, undergo adjuvant CRT/RT. Though Response Evaluation Criteria In Solid Tumours (RECIST 1.1); which relies on single largest dimension for response assessment, is commonly used, it is inadequate for response evaluation of HNSCC.^8,10 RECIST relies on size criteria for response assessment, however, operability of oral cavity tumours and neck dissection post-NACT is determined by shrinkage of tumour from certain sites; hyoid bone (provided tumour does not infiltrate hyoid bone), oedema extending till zygoma, high infratemporal fossa involvement, tumour showing <270° abutment with carotid vessels (for further reduction), in order to get adequate surgical margins, or from vallecula to get better visibility during surgery. Hence, if a tumour shows partial response using RECIST, but does not shrink from the above-mentioned sites, then the tumour still remains inoperable, therefore RECIST is inadequate for response assessment in such cases. So, in these cases, in addition to size criteria, indication for NACT should also be taken into consideration and resolution of tumour from the above-mentioned sites should be considered as a measure of adequate response post-NACT.

RECIST, however, is adequate for assessing response in laryngeal and hypopharyngeal tumours where reduction in size of tumour is the basic intention as further treatment plan is adjuvant RT and not surgery.

Prognostic value of change in tumour volume on post-NACT scan is yet to be explored, however, delineation of accurate tumour margin on post-NACT scan is challenging owing to associated treatment-related changes, hence, definite response assessment in such patients can be made on FDG-PET/CT scan done 12 weeks after adjuvant CRT/RT.⁸Figure 10 shows post-NACT response on MRI in carcinoma tongue.

Immunotherapy

Metastatic and recurrent HNC have been the main target for immunotherapy benefitting 15%-20% of the patients.¹⁴ Various studies are ongoing to channelise the benefit of immunotherapy in neoadjuvant setting, as well as in combination with RT.^11,55–57 Post-immunotherapy response is assessed based on single largest axial dimension using modified RECIST (iRECIST) which takes into account the phenomenon of pseudoprogression seen with immunotherapy.^58–60 Common complications after immunotherapy are pneumonitis and colitis for which imaging might be required. Other post-immunotherapy complications like oral mucositis, autoimmune diabetes, and rash, are diagnosed clinically.¹¹

Post-treatment imaging in HNSCC: recurrence

Tumour recurrence manifests in one or more of the following ways:

Recurrence at the primary tumour site

After radical surgery with or without flap reconstructions, recurrence can manifest at the operated primary tumour bed, surgical or flap margins, or at the under surface of flap, as shown in Figure S2.^21–23 Tumour recurrence deep to flap reconstructions evade detection on clinical examination, hence, the role of imaging in such cases. Earliest sign of recurrence on imaging may be induration of skin, subcutaneous tissue, or fatty component of flap, and progressive soft tissue thickening along with local infiltration.²³ Tumour recurrence appears as enhancing infiltrating mass, mildly hyperdense on non-contrast CT scan and shows intermediate signal intensity on T2WI MRI, with or without bone destruction.²² Bone destruction is best visualized in the bone window setting on CT scan or on non-contrast T1WI.²³ Any new onset nodularity or focal mucosal enhancement should raise the suspicion of recurrence. Oedema has attenuation less than muscle on CT, thus differentiating it from recurrence.²² Post-operative vascularized scar is seen as an enhancing ill-defined soft tissue on CECT and CEMRI, and is of T2 intermediate signal intensity similar to that of recurrence. However, on follow MRI, vascularized scar undergoes retraction and shows T2 hypointensity suggestive of fibrosis, thus, differentiating it from recurrence.^22,23 Multiparametric MRI using diffusion weighted imaging (DWI) and apparent diffusion coefficient (ADC) can differentiate tumour recurrence from post-operative fibrosis and inflammation, as recurrence shows hyperintensity on DWI and signal drop on ADC (diffusion restriction), whereas low ADC value is not seen with post-operative changes.²² One of the studies have shown that an ADC value ≤1.43 × 10⁻³ mm²/s can be used for differentiating locoregional recurrence/residual tumour from post-treatment benign changes.⁶¹

After RT/CRT, focal heterogeneously enhancing infiltrative mass amidst the backdrop of homogeneously enhancing post-therapy fibrosis, should be viewed with suspicion for recurrence on CECT or CEMRI. The presence of enhancing soft tissue mass adjacent to ORN or solid enhancement adjacent to mucosal necrosis should raise the suspicion of recurrence on CECT or CEMRI and biopsy should be performed for confirmation.^22,38 The absence of FDG uptake at the primary site after RT/CRT safely excludes residual disease when performed after 12 weeks.³⁹

If the T2 signal intensity of the soft tissue in the post-treatment scan is of intermediate signal intensity (Figure S3) or similar to that of the tumour in the baseline/treatment naïve scan, then it is highly suggestive of recurrence. Table S2 shows imaging differentiation of recurrence from abscess, seroma, fibrosis, vascularized scar, oedema, and osteoradionecrosis.

Perineural recurrence

Recurrent tumour carries a higher risk of perineural spread, best detected on MRI along with its intracranial extension. Post-operative granulation tissue and scarring can sometimes pose difficulty in interpretation, for which correlation with clinical symptoms might be helpful, as unexplained symptoms of perineural spread appear prior to being evident on images, and follow-up imaging might be required in some cases.^22,62–64 Perineural spread commonly occurs retrogradely towards the brainstem, may show skip lesions, and may further disseminate to leptomeninges, hence, MRI protocol for evaluating perineural recurrence should include entire course (extracranial and intracranial) of the involved nerve along with brain.^64,65 Nerve thickening can occur on first follow-up scan after RT as part of post-treatment change if the nerve is within the field of RT, however, further increase in the nerve thickening with increase in enhancement, widening of skull base foramen, intermediate signal intensity on T2, and diffusion restriction, suggest perineural recurrence as shown in Figure S4.

Neck nodal recurrence

Oral cavity carcinomas usually metastasize to levels I, II, and III nodes, whereas oropharyngeal and supraglottic laryngeal carcinomas frequently metastasize to levels II, III, and IV nodes.²² Carcinomas involving nasopharynx, hypopharynx, and base of tongue usually show metastases to levels II, III, IV, and V nodes. Carcinomas of the nasopharynx, oropharynx, base of the tongue, and supraglottic larynx commonly show bilateral nodal metastases.^6,22 On CECT/CEMRI, metastatic nodes show round shape, loss of fatty hilum, the presence of necrosis, heterogeneous enhancement, and irregular capsule.⁶² Mildly increased enhancement and diffusion restriction (high signal on DWI and signal drop on ADC) are the keys to identifying nodal recurrence amidst post-treatment changes.²² Ultrasonography may be performed followed by fine needle biopsy (FNB), if nodal involvement is indeterminate on cross sectional imaging. FDG-PET/CECT is recommended 10-12 weeks after RT or CRT in case of suspected recurrence or to assess neck response to RT.^2,39,66–68 The absence of FDG uptake at the lymph nodal site after RT/CRT safely excludes residual disease when performed after 12 weeks.³⁹ PET-CT guided active surveillance for locally advanced nodal metastasis (N2/N3) in HNSCC patients treated with radical CRT results in lower costs and complications.⁶⁹

Patients with negative FDG-PET/CECT at first follow-up have better progression free survival (PFS) and OS.^70,71 In follow-up of advanced nodal disease, if there is no FDG uptake or no enlarged node, then it suggests complete response and if intense FDG uptake is detected within normal sized or enlarged node, it is indicative of recurrence.⁶⁸ Mild or no FDG uptake in enlarged node or mild FDG uptake in the normal sized node point towards equivocal findings and warrant ultrasonography guided FNB.⁶⁸

Distant metastasis

FDG-PET/CECT or CECT thorax can detect lung metastasis and can pick up any synchronous malignancies in the upper aerodigestive tract or lungs, which are so common with HNSCC due to tobacco use and alcohol consumption.^2,22,38 CEMRI is the modality of choice for detecting brain metastasis when patient has neurological symptoms.

Recurrence versus second primary

Most of the researchers across the globe are in agreement that a malignant tumour arising from the same site or within 2 cm of the previous tumour, with a disease free interval (DFI) of less than 3 years, is a recurrent tumour. On the other hand, a malignant tumour arising from a separate head and neck cancer subsite, irrespective of the DFI, is a second primary tumour.⁷² Dilemma exists regarding categorization of a tumour into recurrent or second primary when the distance of malignant tumour from the primary site is more than 2 cm and/or the DFI is more than 3 years. Warren and Gates criteria are most widely followed, in which any malignant tumour arising from the same site, irrespective of its distance from the previous tumour, and irrespective of DFI, is a recurrent tumour and not a second primary.⁷³

Post-treatment imaging in HPV positive HNSCC

In head and neck region, HPV positive carcinomas are mostly oropharyngeal in origin.⁶ Only a few tumours of oral cavity, larynx, and paranasal sinuses are HPV related. As per the NCCN guidelines, treatment of HPV positive OPC includes surgical resection along with ipsilateral or contralateral nodal dissection, or definitive RT in the early stages, and CCRT or NACT followed by RT or systemic therapy/RT or surgical resection of the primary and nodal disease in the later stages.⁶ HPV positive OPC have a favourable prognosis with a longer asymptomatic post-treatment period, but with a similar rate of distant metastases as compared to HPV negative OPC (10%-15%).⁷⁴ The pattern of recurrence of HPV positive OPC is however different with more distant and disseminated metastases in multiple atypical organs, unlike HPV negative OPC which present with local recurrence.^74,75 One of the studies have shown that lungs is the commonest site of metastasis.⁷⁴ If NACT is given in HPV positive HNSCC, then CECT/CEMRI can be used for response assessment depending upon the imaging modality used for pre-treatment evaluation. Otherwise, PET-CT is the modality of choice for post-treatment evaluation of HPV positive OPC to rule out local recurrence and distant metastasis after 16 weeks.^74,76 One of the studies have shown that higher risk of distant metastases and lower locoregional failure can be predicted in HPV positive OPC using mid-treatment PET-CT by identifying volumetric nodal metabolic response.^76,77 Most of the distant metastases occur 6 months after treatment completion, hence routine surveillance using FDG-PET/CT may be suggested in HPV positive OPC.⁷⁴ HPV positive OPC usually presents at a younger age and has a better prognosis as compared to HPV negative OPC, hence they are predisposed to acute and long term toxicity from definitive RT/CRT.⁷⁸ Dynamic Imaging Grade of Swallowing Toxicity (DIGEST) score using modified barium swallow imaged with videofluoroscopy can assess radiation induced swallowing inefficiency and dysphagia.^74,79 Increase in mid-therapy T2 SI, decrease in T1SI, and increase in T1 post-contrast MRI throughout the therapy in pharyngeal constrictor muscles can serve as a biomarker for radiation induced dysphagia.

Indications and recommendations on post-treatment imaging

NCCN and ESMO guidelines recommend CECT and/or CEMRI of the primary site within 3-4 months of surgery or primary treatment in patients with locoregionally advanced cancers to serve as a baseline scan prior to commencement of adjuvant therapy.^2,6 Usually, review of the RT planning study is done by the radiologist to rule out any residual disease after surgical resection prior to commencing adjuvant RT, and in case of any concerns, a diagnostic imaging study is performed.

NCCN does not recommend routine imaging surveillance in an asymptomatic treated patient with negative examination if FDG-PET at 3 months has been negative, however, routine annual imaging may be performed for areas difficult to visualize on clinical examination.⁶ Flowchart on the site and treatment specific imaging modality to be used for post-treatment assessment of HNSCC based on NCCN and ESMO guidelines is shown in Figure 11. Both MRI and PET-CECT may be required for follow-up of patients with tumours near or involving the skull base, for evaluation of perineural, intracranial, or intraorbital tumour extension.

Suggested CT and MRI protocols for post-treatment assessment of HNSCC are provided in Table 5.^18,63,80,81 For performing FDG-PET/CECT, 5 MBq/kg body weight is the usual dose of 18FDG and this is combined with CECT.⁸⁰ Standard area of coverage with FDG-PET/CT is skull base to mid-thigh.⁸²

Two consecutively negative FDG-PET/CECT performed within a six-month period has a negative predictive value of 98% and obviates the need for further imaging if clinical signs of recurrence are absent.⁸³ Routine follow-up imaging is not recommended if FDG-PET/CECT done 12 weeks after treatment completion with CRT/RT is negative, unless there is a clinical suspicion of recurrence. Both MRI and PET-CECT should be used for follow-up after CRT and in a suspected case of recurrent nasopharyngeal carcinoma.⁸⁴ When patient has been successfully treated with surgery without the need for adjuvant therapy (no high-risk features on post-operative specimen on histopathology), then no further follow-up imaging is required.⁶

Indications for multidisciplinary tumour board discussions after treatment completion are as follows:

• To confirm recurrence radiologically and pathologically.

• To decide upon feasibility of salvage surgery in case of local recurrence without distant metastasis.

• To decide upon salvage neck dissections in case of regional nodal residual/recurrent disease without distant metastasis.

• To decide upon palliative RT and/or immunotherapy for those with distant recurrent disease.

• To differentiate osteoradionecrosis from recurrence and decide upon further management plan.

Assessment of post-treatment imaging findings: role of Hopkins NI-RADS

Hopkins

Hopkins is a five-point qualitative post-therapy FDG-PET/CECT based response assessment criteria for HNSCC, having excellent negative predictive value and distinguished role in predicting PFS and OS.^85,86,Table S3 shows the Hopkins criteria for post-therapy assessment.⁸⁶

Neck imaging reporting and data systems

American College of Radiology (ACR) developed NI-RADS for comprehensive post-treatment HNSCC and non-squamous HNC (tumours of salivary gland, nasal cavity and paranasal sinuses, orbit and thyroid gland) evaluation and management which includes a lexicon to differentiate benign from malignant post-treatment findings, reporting templates with defined level of suspicion and management recommendations, and post-treatment surveillance imaging, which can be used with CECT with or without FDG-PET and also with MRI.^81,87,88 NI-RADS includes categories 0-4 (incomplete to definite recurrence) on FDG-PET/CECT and MRI, with separate descriptions for primary site and neck in each of them, with their management implications. NI-RADS should be followed while recording imaging findings after definitive treatment completion, and not while the patient is on treatment. Table S4 shows NI-RADS categories on FDG-PET/CECT for primary site and neck node.^87,88 MRI NI-RADS categories for primary tumour are shown in Table S5. CT and MRI of treated HNSCC cases along with their NI-RADS categories are shown in Figure 12 and Figures S5-S9.

Post-treatment surveillance recommendations as per the ACR NI-RADS committee suggests PET-CECT to be performed at 8-12 weeks after completion of definitive treatment and if negative, CECT or PET-CECT to be performed 6 months later. If the first CECT is negative, then only CECT neck should be performed 6 months later and if this second CECT is negative, then CECT neck and chest 12 months later should be done. If two consecutive PET-CECTs are negative, then there is no need for further surveillance imaging.⁸⁷

Advanced imaging techniques for post-treatment HNSCC evaluation

Intravoxel incoherent motion (IVIM), dynamic contrast-enhanced MRI (DCE-MRI), and blood oxygen level dependent (BOLD) MRI are the functional MRI techniques being explored for their role in post-treatment HNSCC evaluation, in addition to DWI.

IVIM non-invasively assesses perfusion and generates quantitative map of small functional blood vessel density, which is a marker of angiogenesis (known to play a role in tumour growth).^89–92 IVIM calculates parameters like pure diffusion coefficient (D), microvascular volume fraction (f), and perfusion-related incoherent microcirculation (D^∗), thus segregating diffusion from perfusion.⁹² Low pre-treatment D and f values and increase in D during treatment entail a favourable response to treatment in HNSCC.⁹³ Differentiation of malignant tumour from CRT related fibrosis was possible using D value with a sensitivity and specificity of 100% in one of the studies.^93,94 One of the studies has shown significant difference in pre-CRT D value between responders and non-responders using IVIM.⁹⁵ In addition, IVIM curtails the cost of contrast agents, acquisition times, and is a boon to those in whom contrast agents are contraindicated. However, IVIM technique, acquisition parameters including b-values, and algorithms for quantitative image analysis need to be standardized.⁹¹

DCE-MRI uses gadolinium contrast to image tissue perfusion. One of the most common parameters derived from DCE-MRI is K^trans, which has shown a strong predictive association with progression free and overall survival in advanced HNSCC.^96–98,K^trans is a volume constant which provides information about the tumour perfusion and vascular permeability. Higher pre-treatment K^trans; mean ± SD K^trans of 0.90 ± 0.54 min⁻¹ and median of 0.88 min⁻¹, from metastatic nodes, has shown to have a better 5-year overall survival in HNSCC as compared to those with a low pre-treatment K^trans.⁹⁷

DCE-MRI has the potential to differentiate between local recurrence and post-treatment change with the help of time-signal intensity (TSI) curves. Plateau or wash-out (type II or III) TSI curves are seen in local recurrence due to recurrent or residual tumour showing early and intense enhancement owing to leaky vessels, whereas, progressive increment (type I) TSI curves are seen in post-treatment change predominantly owing to persistent delayed enhancement in scar, as shown in one of the studies.⁹⁹ DCE-MRI can also predict ORN by demonstrating increased DCE parameters in ORN affected mandible as compared to normal mandible.^74,100

BOLD MRI is a non-invasive technique to evaluate tumour hypoxia, an entity known to reduce effectiveness of CRT and associated with unfavourable outcome in HNSCC.¹⁰¹

Paramagnetic effect of blood deoxyhaemoglobin reduces the signal intensity on T2* images and this forms the basis of identifying tumour hypoxia on BOLD imaging. When oxygen or carbogen is inhaled, there is increase in diamagnetic oxyhaemoglobin which results in increased T2* signal intensity within tumour and the difference in tumour oxygenation is detected by BOLD, which is then used to select patients suitable for anti-hypoxia treatment in radiotherapy.^101,102

Fluoro-misonidazole (¹⁸F-MISO) and ¹⁸F-fluoroazomycin-arabinofuranoside (¹⁸F-FAZA) are the two radiotracers that can be combined with PET to evaluate tumour hypoxia.⁹⁶ One of the studies has shown that tumour hypoxia quantification using the FMISO-PET parameters such as peak tumour-to-background-ratio (TBRpeak); calculated as the standardized uptake value (SUV) at peak divided by the mean SUV within the background volume of interest, with a cut-off 2.0 (P < .001) in week 2, and residual hypoxic volume with 1.6-fold muscle SUV _mean (rHV _1.6) cut-off 0.2, (P < .001), can enable stratification of patients into low (those with a lower than cut-off values indicating lesser tumour hypoxia) and high (those with a higher than cut-off values indicating more tumour hypoxia) risk of resistance to treatment and thus locoregional recurrence.¹⁰³ All the voxels within SUV uptake of FAZA for the gross tumour volume GTV_FAZA-T demonstrating a tumour-to-muscle value equal to or above 1.4 was defined as the hypoxic volume by FAZA PET/CT in one of the studies, and it correlated with poor outcomes in post-RT HNSCC patients.¹⁰⁴

Systematic review of multimodality post-treatment HNSCC studies

Table S6 shows systematic review of multimodality post-treatment HNSCC studies in the last 15 years.^105–109 Pubmed search was conducted for major imaging based post-treatment HNSCC studies in the last 15 years in adult patients >19 years of age reported in English language. Only comparative studies having sensitivity and specificity data were included in the table and those involving single imaging modality were discarded.

Conclusion

Given the shift towards multimodal treatment for head and neck cancers, interpreting post-treatment scan has become complex. Radiologists must possess comprehensive understanding of different surgical techniques and radiation therapy effects to accurately diagnose and manage recurrence and complications. While imaging guidelines vary globally, NCCN recommendations are widely followed.

For locally advanced HNSCC, depending on the specific subsite of the carcinoma, either CEMRI or CECT should be conducted within 3-4 months following surgical resection. This aims to exclude any residual disease/early recurrence before initiating adjuvant CRT/RT. If such a diagnostic scan has not been performed, then typically, the radiologist should review the radiotherapy planning study to ensure the absence of residual disease/early recurrence, and if any concerns arise, further diagnostic imaging studies should be undertaken.

After CRT/RT, it is recommended to utilize FDG-PET/CECT for assessing response 12 weeks post-treatment. If the results are negative, routine imaging surveillance is unnecessary unless recurrence is suspected. In cases where recurrent nasopharyngeal carcinoma is suspected, both MRI and PET-CECT should be employed for follow-up evaluation.

In the event of intense FDG uptake detected within a normal sized or enlarged node on follow-up PET-CECT, an ultrasound-guided biopsy is recommended to confirm recurrence.

If a patient has undergone successful surgery without the necessity for adjuvant therapy, meaning there are no high-risk features on the post-operative histopathology specimen, further follow-up imaging is unnecessary.

Annual imaging on a routine basis may be considered for regions that are challenging to visualize during clinical examination.

Following completion of treatment, it is customary to conduct multidisciplinary tumour board discussions. These discussions serve several purposes, including confirming recurrence, determining the necessity for salvage surgeries and neck dissections in instances of local recurrence or residual/regrown regional disease without distant metastasis, devising palliative therapies for distant recurrence, and distinguishing between osteoradionecrosis and recurrence to strategize further management.

In HPV positive oropharyngeal squamous cell carcinoma, routine surveillance with FDG-PET/CT may be contemplated. This is because the majority of distant metastases tend to manifest within 6 months following completion of treatment.

NI-RADS guidelines should be adhered to for documenting post-treatment imaging findings after completion of treatment for both squamous cell carcinomas of the head and neck and non-squamous cell cancers.

There is a necessity to prospectively validate the efficacy of advanced imaging techniques such as functional MRI, IVIM, and BOLD imaging. This validation should involve a larger sample size to ascertain their potential utility for predicting response to chemoradiotherapy.

Author contributions

(1) Guarantor of integrity of the entire study: Nivedita Chakrabarty, Abhishek Mahajan, (2) Study concepts and design: Nivedita Chakrabarty, Abhishek Mahajan, (3) Literature research: Nivedita Chakrabarty, Abhishek Mahajan, (4) Manuscript preparation: Nivedita Chakrabarty, Abhishek Mahajan, and (5) Manuscript editing: Nivedita Chakrabarty, Abhishek Mahajan, Archi Agrawal, Kumar Prabhash, Anil K. D’Cruz.

Supplementary material

Supplementary material is available at BJR online.

Funding

This article did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflicts of interest

The authors declare that there is no conflict of interest.

Ethics statement

As this is a review article, Institutional Ethics Committee approval was not required and patient identity has not been disclosed on any image.

References