Prognostic role of PET/CT volumetric parameters in paediatric Hodgkin lymphoma: A systematic review and expert recommendations from the International SEARCH for CAYAHL Group

Summary

In adults with classic Hodgkin lymphoma (cHL), metabolic tumour volume (MTV) is valuable for risk stratification. This systematic review explored the role of volumetric parameters in children, adolescents and young adults (CAYA) with cHL. This systematic review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) method. Five databases were searched on 16 October 2024. Eligible studies were peer‐reviewed manuscripts, written in English, published since database inception that included patients ≤21 years of age with a diagnosis of cHL treated with chemotherapy ± radiotherapy with a baseline ¹⁸F‐Fluorodeoxyglucose (F‐FDG) PET/CT scan for MTV quantification. The search strategy identified 3669 studies, of which 35 were eligible for inclusion. Fifteen studies (43%) evaluated a cohort exclusively ≤21 years of age. All eight studies that explored the association of baseline PET parameters with known risk factors identified a significant correlation. Nine of 12 papers (75%) demonstrated a statistically significant correlation with disease response to therapy on PET/CT. Fifteen of 20 papers (75%) demonstrated an independent statistically significant correlation with clinical outcomes. In the CAYAHL population, volumetric PET parameters are valuable for risk stratification despite heterogeneity in results. Future harmonization of the segmentation approach will facilitate incorporation of MTV into clinical practice.

A visual overview of the key findings from the 35 papers included in this systematic review evaluating the use of PET volumetric parameters in children, adolescents and young adults with Hodgkin lymphoma.

INTRODUCTION

Hodgkin lymphoma is a metabolically active tumour, and ¹⁸F‐Fluorodeoxyglucose (F‐FDG) PET/CT is the current standard of care for staging evaluation and response assessment in patients of all ages. Staging based on the Ann Arbor classification system depends on accurate evaluation of the initial disease extent, including the detection of bulky adenopathy. PET/CT is superior to CT alone for accurate staging of classic Hodgkin lymphoma (cHL). ¹ , ² Findings on PET scans conducted not only at baseline but also after chemotherapy hold significant prognostic value. The detection of residual FDG‐avid disease after chemotherapy has been associated with an increased risk of disease progression. ³ Therefore, disease response on interim PET scans, classified according to the Deauville criteria, has become a key factor in guiding subsequent management, with treatment intensified for patients with positive interim scans or de‐escalated for those with negative scans. ⁴ , ⁵ , ⁶ , ⁷

Metabolic tumour volume (MTV), defined as the sum of all voxels with an elevated standardised uptake value (SUV) within tumour lesions, reflects the amount of viable tumour in the body and serves as a quantitative indicator of disease burden. According to the adult literature, baseline MTV is independently prognostic, often outperforming traditional clinical risk factors and staging systems. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ These data, demonstrating that MTV is a robust, reproducible biomarker that enhances risk stratification in adult cHL, support its integration into future risk‐adapted treatment algorithms.

The role of MTV in CAYAHL is unclear. Factors unique to children—including variability in body size and frequent reactive lymphadenopathy or activated brown adipose tissue—may affect PET scan interpretation, SUV measurement and tumour segmentation. Therefore, PET data from adults may not apply to paediatric patients, and the role of MTV must be assessed in this population specifically. This systematic review evaluated the literature regarding the use of baseline volumetric PET parameters in children, adolescents and young adults (CAYA) with cHL.

The Staging, Evaluation and Response Criteria Harmonization for Childhood, Adolescent, and Young Adult Hodgkin Lymphoma (SEARCH for CAYAHL) working group comprises an international collaboration of experts in paediatric oncology, radiology, nuclear medicine and radiation oncology. Established in 2011, the group aims to develop a unified framework for staging and response assessments across paediatric Hodgkin lymphoma clinical trial groups. This study by SEARCH for CAYAHL was designed to address two questions: (1) Do data support a role for MTV in risk stratification in CAYAHL? (2) What are the next steps to support the integration of MTV into clinical practice?

METHODS

The systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) method. ¹⁴ The search strategy, based on the PICO principle, included the CAYAHL population with baseline disease evaluation on an ¹⁸F‐FDG PET/CT (Table S1). Five databases were searched on 16 October 2024: MEDLINE (Ovid), EMBASE (Ovid), CINAHL (EBSCO), EBM Reviews (OVID) and the Web of Science Core Collection. Eligible studies imported into Covidence for further analysis included peer‐reviewed manuscripts in English, published from the inception of the database until the search date. These studies involved patients aged 21 years or younger, either as a dedicated paediatric cohort or combined with adults, diagnosed with cHL and assessed using ¹⁸F‐FDG PET/CT scans with volumetric parameters calculated (Table S2).

RESULTS

Study characteristics

The search strategy identified 3669 studies. Duplicate screening removed 1085 articles. Two thousand five hundred and three articles were removed after two‐person title and abstract screening. Of 81 papers retrieved for full‐text review, 35 met the inclusion criteria for further analysis and data extraction, which was performed independently by two reviewers (Figure 1). Fifteen studies (43%) included a dedicated CAYA cohort, ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ and 20 (57%) evaluated a combination of adult and paediatric participants. ¹⁰ , ¹¹ , ¹² , ¹³ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ Twenty‐nine papers (83%) assessed for an association of PET volumetric parameters with clinical data ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³³ , ³⁴ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴³ , ⁴⁴ , ⁴⁵ ; the other six studies focused specifically on technical features of disease segmentation ²⁶ , ²⁷ , ³² , ³⁵ , ⁴¹ , ⁴² (Table S3).

FIGURE 1.

Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) flow diagram detailing the systematic review process.

Patient, disease and treatment data

Most cohorts included more males than females, and this difference was greater in the dedicated CAYA study populations. The female‐to‐male median and mean ratios were 1:1.55 and 1:1.9 in CAYAHL, ¹⁵ , ¹⁶ , ¹⁷ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ compared with 1:1.2 and 1:1.4 in the mixed age cohorts ¹⁰ , ¹¹ , ¹² , ¹³ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ (Table S2). The cohorts included patients with both early‐ and advanced‐stage disease in 24 studies (69%), ¹² , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁸ , ²⁹ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ early‐stage disease only in 5 (14%) studies, ¹⁰ , ¹¹ , ³⁰ , ³¹ , ³² advanced‐stage disease only in 3 (9%) studies ¹⁵ , ²⁰ , ²⁷ and relapsed/refractory disease in 2 (6%) studies. ¹³ , ³³ Patients were treated with chemotherapy with or without radiotherapy according to local practice or clinical trial stipulations. Immunotherapy was not identified as a treatment strategy in any of the evaluated papers.

Volumetric PET parameters

Twenty‐nine papers assessed for an association between baseline PET parameters and established prognostic factors, PET/CT‐based disease response assessments and patient outcomes.

Established risk factors

Eight of 29 papers (28%) assessed the association of baseline metabolic parameters with known prognostic factors, demonstrating statistical significance for bulky disease, Ann Arbor stage and treatment group/level. ¹⁰ , ¹¹ , ¹² , ¹⁹ , ²³ , ²⁵ , ³⁰ , ³⁷ Baseline PET parameters added value to the predictive ability of patient classification by treatment group/level, ¹¹ could substratify early‐stage unfavourable disease into low‐ and high‐risk groups ¹⁰ , ¹¹ and were superior to stage or response assessments at predicting the risk of disease relapse after upfront therapy. ³⁷

PET/CT response

Twelve of 29 papers (41%) assessed the association of baseline PET volumetric parameters with disease response on post‐chemotherapy PET/CT scans, most commonly interpreted using the Lugano criteria. ¹⁰ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁸ , ³⁴ A statistically significant association was identified in nine of these 12 papers (75%). ¹⁰ , ¹⁵ , ¹⁷ , ¹⁸ , ²⁰ , ²¹ , ²³ , ²⁴ , ²⁸ Of seven studies that confirmed the value of baseline PET parameters to predict response on an interim PET/CT, ¹⁰ , ¹⁵ , ¹⁷ , ¹⁸ , ²⁰ , ²³ , ²⁸ six evaluated a paediatric cohort treated with two cycles of standard chemotherapy. ¹⁵ , ¹⁷ , ¹⁸ , ²⁰ , ²³ , ²⁸ One paper found total lesion glycolysis (TLG = MTV × SUVmean), but not MTV, to be predictive. ²⁰ Four papers reported a statistically significant association of baseline PET parameters with disease response on a PET/CT performed at the end of therapy. ¹⁷ , ²¹ , ²⁴ , ³⁴ Four papers explored and demonstrated that a combination of baseline PET parameters with interim PET/CT response improved the predictive value of PET/CT ¹⁰ , ¹² , ¹³ , ²⁰ (Table 1).

TABLE 1.

Studies evaluating PET parameters with disease response on PET/CT. [Colour table can be viewed at wileyonlinelibrary.com]

Study	Disease response assessment on PET/CT		Statistical significance of baseline PET parameter to predict response	Treatment	Metabolic PET parameter
	Interim PET/CT	End of treatment PET/CT
Cottereau, 2018	Yes	N/A	Yes	H10 trial. 2c ABVD	MTV
Georgi, 2018 a	Yes	N/A	Yes	EuroNet‐pHL‐C1. 2c OEPA	MTV
Hussien, 2015 a	Yes	N/A	No	GPOH/EuroNet‐pHL‐C1. 2c OEPA	MTV/TLG
Lopci, 2022 a	Yes	Yes	Yes	EuroNet‐pHL‐C2. 2c OEPA 2c or 4c DECOPDAC or COPDAC± RT	MTV/TLG
Lopci, 2023 a	Yes	N/A	Yes	EuroNet‐pHL‐C2. 2c OEPA	MTV/TLG
Milgrom, 2023 a	Yes	N/A	Yes	AHOD0831. 2c ABVE‐PC	TLG
Naguib, 2021	N/A	Yes	No	Not detailed	MTV/TLG
Reed, 2021 a	N/A	Yes	Yes	6c ABVD	MTV
Rodriguez‐Taroco, 2021 a	Yes	N/A	No	1‐LH‐GHOP 2005 protocol ABVE +RT (64%) BEACOPP + RT (100%)	MTV/TLG
Rogasch, 2018 a	Yes	N/A	Yes	EuroNet‐pHL‐C1. 2c OEPA	MTV
Sharma, 2012 a	N/A	Yes	Yes	Not detailed	TLG
Yadgarov, 2024 a	Yes	Yes	Yes	EuroNet‐PHL‐C1/GPOH 2c OEPA	MTV/TLG

Abbreviations: HL, Hodgkin lymphoma; MTV, metabolic tumour volume; N/A, not assessed; RT, radiotherapy.

^a Study used a CAYAHL cohort exclusively; all others included paediatric and adult patients with HL. Chemotherapy abbreviations taken from the paper. Blue highlight = those studies which did not find a positive correlation between baseline PET parameter and response assessment on PET/CT.

Patient outcomes

Twenty of 29 studies (69%) correlated baseline volumetric PET parameters with oncologic outcomes. ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁹ , ²⁰ , ²¹ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³³ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴³ , ⁴⁴ , ⁴⁵ Fifteen of 20 papers (75%) demonstrated a statistically significant association of baseline PET parameters with progression‐free survival (PFS) and/or overall survival (OS) and/or refractory disease. ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁹ , ²⁰ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³³ , ³⁶ , ³⁷ , ³⁸ , ³⁹ Four of the 15 studies found that TLG was prognostic, rather than MTV. ²⁰ , ²⁹ , ³⁰ , ³⁷ Three papers did not show statistical significance on multivariate analysis, all of which evaluated a mixed‐stage cohort ²¹ , ²² , ⁴⁴ (Table 2).

TABLE 2.

Studies evaluating baseline PET parameters with survival outcomes. [Colour table can be viewed at wileyonlinelibrary.com]

Study	Length of follow‐up (months)	Patient outcomes evaluated after a period of clinical follow‐up			Statistically significant as an independent prognostic factor	Treatment	Metabolic PET parameter
		Progression‐free survival	Overall survival	Other
Akhtari, 2018	59.5	Yes	Yes	—	Yes	ABVD 1C‐5C in 89% and IFRT in 70%	MTV/TLG
Alagoz, 2017	73	Yes	N/A	—	Yes	Protocol ABVD +30Gy RT	TLG
Cottereau, 2018	55	Yes	Yes	—	Yes	Standard arm ABVD 3C(F) 4C(U) +30Gy INRT	MTV
Frood, 2022	24	Yes	N/A	—	Yes	Standard ABVD ±RT not detailed	MTV
Herraez, 2021	45	Yes	Yes	—	Yes	I/II = 2‐4C ABVD ± RT. III/IV = 6C ABVD	TLG
Kahle, 2019	42	Yes	N/A	—	Yes	Standard ABVD + RT or escBEACOPP	MTV/TLG
Kanoun, 2014	50	Yes	Yes	—	Yes	Standard LYSA Anthracycline‐based CT. I/II 4‐6C + 20‐36Gy. III/IV 8C	MTV
Kanoun, 2015	39	MTV	N/A	—	Yes	Standard LYSA Anthracycline‐based CT. I/II 4‐6C + 20‐36Gy. III/IV 8C	MTV
Lue, 2019	48	Yes	Yes	N/a	No	Anthracycline‐based (55%) + RT (45%)	MTV/TLG
Milgrom, 2019	1.5	N/A	N/A	Refractory	Yes	ABVD 2C‐6C 98% + RT (70%)	MTV
Milgrom, 2021 a	79.4	Yes	N/A	—	Yes	ABVE‐PC(2C) + RER ABVE‐PC (2C) ± IFRT 21Gy SER 2C ABVE‐PC ± 2C DECA + IFRT	MTV
Milgrom, 2023 a	49	TLG	N/A	—	Yes	ABVE‐PC (2C); RER 2C ABVE‐PC or SER also had 2C ifosfamide _ vinorelbine ±or 21Gy/14 IFRT	TLG
Moskowitz, 2017	40	Yes	N/A	—	Yes	Brentuximab vedotin 2C. ±ICE 2C. ± RT. ASCT.	MTV
Reed, 2021 a	20	Yes	Yes	—	No	6c ABVD	MTV/TLG
Santos, 2024	81.6	Yes	Yes	CIF	No	Standard GHSG protocol ABVD (2‐6C) ±RT	MTV/TLG
Tseng, 2012	50	Yes	Yes	—	No	Varied Stanford V, ABVD, VAMP, Beacopp ±RT 20, 25.5, 36Gy	MTV
Voorhees, 2022	N/A	Yes	N/A	—	Yes	Bridging chemotherapy ± CD30 CAR‐T cell therapy ± ASCT	MTV
Yadgarov, 2024 a	35	Yes	N/A	—	Yes	EuroNet‐PHL‐C1/GPOH	MTV/TLG
Zhou, 2020 a	36	Yes	N/A	—	Yes	Not detailed	TLG
Zhou, 2021	40	Yes	N/A	—	No	Low risk ABVD Inter = ABVD (4‐6cy) + IFRT High = ABVD (6‐8cy) ±RT	MTV/TLG

Abbreviations: CIF, cumulative incidence of failure; MTV, metabolic tumour volume; N/A, not assessed; RT, radiotherapy.

^a Study used a CAYAHL cohort exclusively; all others included paediatric and adult patients with HL. Chemotherapy abbreviations taken from the paper. Green highlight = those studies which did not find a positive correlation between baseline PET parameter and patient oncologic outcomes.

Seventeen of 29 studies (59%) reported a proposed cut‐off value for the chosen PET parameter. ¹⁰ , ¹¹ , ¹² , ¹³ , ²⁰ , ²² , ²³ , ²⁴ , ²⁸ , ²⁹ , ³⁰ , ³³ , ³⁷ , ³⁹ , ⁴⁰ , ⁴³ , ⁴⁵ The majority considered mixed stages of disease together and reported a single cut‐off value for MTV (for all papers: cut‐off median was 225 mL and mean was 227 mL).

Six papers explored higher order radiomic features, ²² , ²³ , ³¹ , ³⁶ , ⁴⁰ , ⁴⁵ and four studies concluded that textural features outperformed MTV/TLG as biomarkers of disease alone or when incorporated into a classification model. ²² , ³¹ , ⁴⁰ , ⁴⁵ Three papers chose features from a Grey Level Co‐Occurrence Matrix ²² , ³¹ , ⁴⁵ ; however, the features that performed well were not consistent between studies. Dmax, a measure of disease spread, was found to be an independent prognostic marker for PFS by one group. ⁴⁵

Paediatric population considerations

Two author groups considered that the wide range in body size within the CAYA population would influence the relative metabolic volume of disease. ¹⁶ , ²⁵ It remains unclear whether SUVmax calculated by body weight, lean body mass or body surface area is the preferred method for children, as the two studies had differing results.

Technical factors

Segmentation

Thirty‐four of 35 (97%) studies documented the segmentation approach. Eleven of 34 studies concluded the SUV2.5 absolute fixed threshold performed best. ¹¹ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ²⁰ , ²⁹ , ³¹ , ³² , ³⁸ , ³⁹ One study demonstrated that SUV4.0 was more robust than an adaptive method. ³⁶ The relative fixed threshold set at 41% SUVmax as recommended by the European Association Nuclear Medicine (EANM) was favoured in 10 papers. ¹⁰ , ¹² , ¹³ , ²² , ³³ , ³⁷ , ⁴⁰ , ⁴¹ , ⁴³ , ⁴⁵ On direct comparison, Yadgarov et al. reported that 41% SUVmax and SUV2.5 were equally effective at predicting patient outcomes in CAYAHL. ²⁸

To delineate disease, three studies evaluated an AI tool that was trained using convolutional neural networks, ²⁶ , ²⁷ , ⁴² and one paper relied on a semi‐automatic method. ³⁵ Compared to physician ground truth segmentation, all studies confirmed a role for these tools in tumour delineation, providing rapid, reproducible and reliable segmentation with reduced physician input (Table 3).

TABLE 3.

Segmentation methods reported per study. [Colour table can be viewed at wileyonlinelibrary.com]

Study	Type of study	Preferred method	Other methods
Akhtari, 2018	Clinical	SUV2.5	Manual
Alagoz, 2017	Clinical	42% SUVmax
Camacho, 2020	Technical	Semi‐automatic	20% SUVmax and 40% SUVmax
Cottereau, 2018	Clinical	41% SUVmax
Frood, 2022	Clinical	SUV4.0	SUV4.0 and adaptive using 1.5× SUVmean Liver
Georgi, 2018 a	Clinical	SUV2.5
Herraez, 2021	Clinical	41% SUVmax
Hussien, 2015 a	Clinical	SUV2.5
Kahle, 2019	Clinical	Combined	SUV2.5 + (SUVmean Liver +2SD)
Kanoun, 2014	Clinical	41% SUVmax
Kanoun, 2015	Clinical	SUV2.5	SUV2.5, 41% SUVmax, 125% liver SUVmax, 140% liver SUVmax
Lopci, 2022 a	Clinical	SUV2.5	SUV2.5, 41% SUVmax, adaptive (Liver background)
Lopci, 2023 a	Clinical	SUV2.5
Lue, 2019	Clinical	41% SUVmax
Martin‐Saladich, 2020	Technical	SUV2.5	SUV2.5, 41% SUVmax, 50% SUVmax
Meignan, 2014	Technical	41% SUVmax
Milgrom, 2019	Clinical	SUV2.5
Milgrom, 2021 a	Clinical	SUV2xBP	%SUVmax at 20, 30, 40, 50, 60%, SUV2.5, SUV3.0, 1.5xLvmean, 1.5xSUVmean Liver +2SD, 2SUVmean Blood Pool
Milgrom, 2023 a	Clinical	SUV2.5	SUV2.5 and 40% SUVmax
Moskowitz, 2017	Clinical	41% SUVmax
Naguib, 2021	Clinical	Not detailed
Reed, 2021 a	Clinical	Combined	SUV2.5 and 41% SUVmax
Rodriguez‐Taroco, 2021 a	Clinical	41% SUVmax
Rogasch, 2018 a	Clinical	Adaptive bkg	SUV2.5, 41% SUVmax, Background adaptive method
Sadik, 2024	Technical	AI tool	Manual and software
Santos, 2024	Clinical	41% SUVmax
Sharma, 2012 a	Clinical	% SUVmax	Variable 40%, 60%, 70% depending on size
Tatci, 2017 a	Clinical	40% SUVmax
Tie, 2024 a	Technical	AI tool	SUV2.5 + 40% SUVmax
Tseng, 2012	Clinical	Adaptive	50% cut‐off value in a region‐growing procedure to SUVmax of tumour in a 1‐cm radius
Voorhees, 2022	Clinical	41% SUVmax
Weisman, 2020 a	Technical	AI tool	SUV2.5 + 40% SUVmax
Yadgarov, 2024 a	Clinical	2 methods equal	SUV2.5 and 41% SUVmax
Zhou, 2020 a	Clinical	SUV2.5
Zhou, 2021	Clinical	41% SUVmax

^a Study used a CAYAHL cohort exclusively; all others included paediatric and adult patients with HL. Blue highlight = those studies which used 41% SUVmax. Green highlight = study which used SUV4.0. Yellow highlight = studies which used an AI tool or semi‐automatic method based on convolutional neural networks. Several papers detailed the use of more than one software package within the cohort.

Software

Twelve different software packages were used for tumour segmentation and quantification of volumetric PET parameters (Table S4).

DISCUSSION

To the best of our knowledge, this work represents the first systematic review to explore the prognostic value of PET volumetric parameters in CAYAHL. It is critical to address this question in the CAYA population specifically. Although the normal biodistribution of FDG is consistent regardless of patient age, certain factors unique to children may affect scan interpretation, SUV measurement and tumour segmentation (e.g. reactive lymphadenopathy, brown adipose tissue activation, thymic rebound hyperplasia); therefore, data from adults may not apply to CAYAHL. Thus, this work addresses a data gap.

Data support a role for baseline MTV in risk stratification in CAYAHL. Among 35 papers that included CAYAHL, most showed that high baseline MTV was associated with adverse baseline prognostic factors, unfavourable response to chemotherapy on PET/CT and inferior PFS/OS. ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁵ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²³ , ²⁴ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³³ , ³⁶ , ³⁷ , ³⁸ , ³⁹ The predictive value of the baseline MTV may be enhanced when combined with the interim PET response. ¹⁰ , ¹² , ¹³ , ²⁰ This combined approach has been utilized to adapt therapy in the adult RAFTING trial (NCT04866654) and may play a role in future trials in the CAYAHL population. ⁴⁶ The added value of higher order radiomic features was shown by a small number of studies ²² , ³⁶ , ⁴⁰ , ⁴⁵ ; however, additional research is needed to elucidate their clinical utility.

Conversely, a minority of studies did not identify a statistically significant association between baseline PET volumetric parameters and clinical outcomes. Many of these studies used unique segmentation approaches, which we hypothesize contribute to the discrepant findings. ¹⁶ , ³⁴ , ⁴⁴ This observation highlights the need to standardize segmentation methods. Until recently, there was no common segmentation methodology for lymphomas. ⁴⁷ , ⁴⁸ However, a recent publication by Boellaard et al. provides a reproducible benchmark dataset that can be used as a reference standard for segmentation in adult lymphomas. ⁴⁹ While the limitations of the test set are documented, this paper marks a milestone in the standardization of metabolic tumour burden quantification in adults. ⁴⁹ , ⁵⁰ Similarly, a standardized segmentation approach is necessary in the CAYAHL population.

Another clinically important aim is to identify an MTV cut‐off value that reliably discriminates patients with a higher risk of relapsed/refractory disease. Current studies are not conclusive because cut‐off values have been identified within various subgroups that have received heterogenous therapies. Cut‐off values may be useful to define not only high total‐body disease burden but also bulk of individual nodal masses. Current approaches for defining bulk vary between the CAYA population and adults, with bulky mediastinal adenopathy characterized radiographically by a ratio on chest radiographs in children and by axial measurements on CT scans in adults, or using a volumetric measure on CT of ≥200 mL. Furthermore, subjectivity exists regarding the best plane in which to measure extra‐mediastinal bulk. Using a standardized method, MTV offers a more reliable and objective assessment of bulk, minimizing variations due to measurement techniques, patient age and inter‐observer differences. Future work is required to define useful values for risk stratification across cohorts.

For MTV to be used as a part of clinical practice, segmentation must be rapid and reliable. An AI segmentation tool is being developed, refined and validated utilizing paediatric trial datasets. ²⁶ , ²⁷ The use of AI may enhance the ease and reproducibility of tumour segmentation and thus may contribute to the implementation of PET volumetric parameters into clinical practice.

Bias in studies

Heterogeneity exists among the cohorts; 57% of papers evaluated a cohort of mixed ages ¹⁰ , ¹¹ , ¹² , ¹³ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ and 71% of mixed disease stages. ¹² , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁸ , ²⁹ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ Twenty‐six studies evaluated an institutional cohort ¹⁶ , ²¹ , ²² , ²⁴ , ²⁵ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ (23/26 single site), while eight studies used data from multi‐institutional paediatric cHL trials. ¹⁵ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²³ , ²⁶ , ²⁷ Nine studies used a prospective design, ¹⁶ , ¹⁷ , ¹⁸ , ²² , ³³ , ³⁴ , ³⁵ , ⁴¹ , ⁴² while 26 were retrospective. Treatment strategies varied across studies, with some cohorts receiving chemotherapy as early as 2003 with agents that are no longer standard of care, differing institutional treatment protocols or trial requirements and inconsistency in the use of radiotherapy. None of the evaluated studies utilized immunotherapy. Often, patients with a higher baseline disease burden were selected for intensified systemic therapy and/or radiotherapy, which could influence the findings. Overall, it has not been possible to delineate the effect of treatment with chemotherapy ± radiotherapy on patient outcomes related to baseline MTV given the degree of heterogeneity in study cohorts. Hodgkin lymphoma occurs more frequently in males than females, with a larger gender disparity within the CAYA population; however, there are no available data regarding the influence of gender on the relationship between baseline volumetric PET parameters and oncologic outcomes. Heterogeneity in parameter calculation results from technical differences in the software, segmentation methods, PET/CT equipment, quality control, image acquisition and reconstruction algorithms at different sites.

Next steps

Our proposed next steps involve collaboration across study groups to establish a clear and consistent protocol for measuring MTV in CAYAHL, establish MTV cut‐off value(s) for risk stratification, evaluate if the combination of baseline MTV and interim PET/CT response provides better risk stratification than either parameter alone, explore the association of MTV with other biomarkers and develop a paediatric risk score similar to the metabolic prognostic index.

Standardization of the segmentation approach will require validation of the absolute fixed threshold approach using SUV2.5, exploration of the absolute fixed threshold approach using SUV4.0 as per the adult benchmark dataset and refinement of an AI tool to automate processes. Further research must explore the effect of body size relative to MTV for prognosis. Additionally, it must account for the intensity of therapy that was given and explore the impact of novel systemic agents. These data can be derived from currently ongoing studies, such as COG AHOD2131 (NCT05675410), and archived imaging data from clinical trials (Table S5). Ultimately, we aim to develop a paediatric risk score that incorporates PET parameters along with established and novel clinical and laboratory metrics (e.g. Ebstein Barr Virus (EBV) status, thymus‐ and activation‐regulated chemokine (TARC), Programmed Death Ligand 1 (PDL1) expression and circulating tumour DNA); such a score may direct risk‐adapted therapy on a prospective clinical trial in CAYAHL, thus advancing management of CAYAHL.

CONCLUSION

A thorough and robust systematic review of the literature has shown that metabolic tumour volume (MTV) correlates with current methods for risk stratification, predicts response to treatment on PET/CT and is an independent prognostic factor in CAYAHL. The heterogeneity in PET quantification techniques underscores the importance of standardizing MTV measurements. The proposed next steps aim to facilitate adoption of MTV as a key tool for risk stratification and treatment selection, with the overarching goal of optimizing the care of children, adolescents and young adults with Hodgkin lymphoma.

AUTHOR CONTRIBUTIONS

This manuscript is a collaborative effort between all 23 co‐authors who are members of the SEARCH for CAYAHL group. All co‐authors have contributed to the conception and design of the work. S. Claire Gowdy and Sarah A. Milgrom have acquired and analysed the data. All co‐authors have contributed to the interpretation of the data. S. Claire Gowdy has drafted the manuscript, and all co‐authors have critically appraised the work, suggested revisions and enhanced the intellectual content through review and group discussion. All co‐authors have approved the final content of the manuscript.

FUNDING INFORMATION

The authors have nothing to report.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to disclose.

Supporting information

Data S1.

DATA AVAILABILITY STATEMENT

Retrospective review of manuscripts is freely available.