Current Status of Revisions to the Lugano Classification in Lymphoma

ABSTRACT

In the decade since publication of the Lugano Classification (Cheson et al, J Clin Oncol 2014,32:3059–3068; Barrington et al, J Clin Oncol 2014, 32:3048–3058), major advances in lymphoma therapy and assessment, including metabolic tumor volume (MTV) and circulating tumor DNA (ctDNA) prompted a workshop at the International Conference on Malignant Lymphoma‐17 entitled “Lugano Classification: Looking Toward the Future”, to determine whether a revision was warranted and what it should look like. This report summarizes the conclusions of that workshop and a subsequent questionnaire of the participants. It was concluded that, whereas, minor modifications could be made now, the current classification should remain the standard until the clinical benefit of MTV and ctDNA are firmly established and practical considerations demonstrated. An ICML sponsored standing committee will monitor progress and ensure that a revision of the Lugano Classification will be proposed when warranted.

1. Introduction

The current four stage anatomic system for lymphoma was first developed to assist radiation oncologists in focusing their therapy, or, alternatively, identifying patients unsuitable for radiation, as well as enabling meaningful comparisons amongst different data sets [1, 2]. Over the ensuing years, major advances in therapy and technology have driven modifications in how staging is conducted and response measured. With the availability of CT scans, staging laparotomy, splenectomy and lymphangiography were eschewed for non‐invasive approaches [3]. PET scans revolutionized staging and response evaluation, and were acknowledged in the NCI‐Working Group criteria [4], which focused primarily on large B‐cell lymphoma (LBCL) and Hodgkin lymphoma (HL). Additional information of the utility of PET‐CT, resulted in the 2007 International Working Group modification [5]. Standardization with the 5‐point Deauville Score (DS) and the development of metabolic response categories for assessment of FDG‐avid histologies by PET‐CT resulted in the Lugano Classification (LC) of 2014 for adults with nodal lymphomas, which remains the standard to the present time [6, 7].

However, it has now been more than a decade since the previous LC, and major advances have been made in our understanding of the genetics and biology of lymphomas, notably ctDNA [8], as well as in imaging techniques including radiomics, specifically the use of metabolic tumor volume (MTV) [9]. Additionally, with increasing collaboration in adolescent/young adult clinical trials, inclusion of pediatric criteria is essential to allow for use of the LC in all patients. Thus, a workshop was held at the International Conference on Malignant Lymphomas (ICML)‐17 entitled “Lugano Classification: Looking Toward the Future”, to determine whether a revision was warranted and what is should look like.

2. Methods

In the months prior to the workshop experts were recruited into five subcommittees: Nodal Lymphomas, Extranodal Lymphomas, Imaging, Molecular/genetics, and Pediatrics. Each subcommittee met several times virtually to identify those relevant issues for presentation and discussion at the ICML workshop.

The 80 international participants included oncologists, hematologists, pathologists, imaging specialists, radiation oncologists, and others interested in lymphoma genetics and biology. Unlike prior classifications, discussions regarding pediatric and extranodal lymphomas were included.

Following the workshop, a questionnaire was developed and submitted to the participants to assess their impressions about critical issues discussed at that workshop. Responses were categorized as “Strongly Agree”, “Agree”, “Neutral”, “Disagree”, “Strongly Disagree”, “Not Qualified”. Free text comments were allowed. Seventy attendees responded, and their responses analyzed (Appendix 1). There has been considerable interest in the deliberations of that Workshop and its implications for a revision of the LC. What follows is a summary of the meeting discussions and the questionnaire results; it is not intended to provide specific recommendations for patient management.

3. Results and Discussion

3.1. Replacing/Revising the LC

The overriding question addressed was whether it is time for a revision of the LC. While the majority expressed that the LC would benefit from updating, no specific recommendations were offered. MTV is emerging as a strong prognostic biomarker [10], and ctDNA shows great promise for staging, prognosis, response assessment, and monitoring [8]. However, the conclusion was that these new approaches are not yet widely available and ctDNA is not standardized nor affordable to lymphoma physicians around the world for use in general practice. Other major issues are discussed below.

3.2. Staging

A primary consideration was what should be the current definition of “Staging”? Should it remain simply the identification of anatomic sites of disease, or should it be modernized to include pathologic, molecular/genetic, and imaging/radiomic features that would identify subsets of patients with varying outcomes requiring different therapeutic approaches? Although the majority of participants agreed that the same staging classification should be used for all nodal lymphomas, there was considerable controversy as to whether the current historical anatomic four‐stage Ann Arbor/Lugano staging system should remain as is or be enhanced or even supplanted by one with fewer stages that is more consistent with how patients are actually managed. However, there was no agreement as to what the structure of a new system should be. One option proposed was a 2‐stage system (limited stage—stages I and II without poor risk features; advanced—stages I and II with poor risk features, stages III and IV). A three‐stage system was also suggested. No conclusions were drawn and further discussions will be ongoing. Specific recommendations should be developed regarding staging and restaging of extranodal lymphomas.

Nonetheless, the consensus was unanimous recommending the continued use of PET‐CT for staging and response assessment of HL, aggressive B‐cell and T‐cell lymphomas, follicular and mantle cell lymphoma. Contrast‐enhanced CT (ceCT) is generally not useful for routine staging and response assessment of FDG‐avid lymphomas and is discouraged in general clinical practice and should not be included in guidelines. Exceptions include when ceCT is required for radiation planning or other distinct indications (e.g., assessment of mediastinal thrombosis), or if a clinical emergency precludes waiting for a PET‐CT scan (e.g., Burkitt lymphoma). Given the variable FDG‐avidity in marginal zone lymphoma according to subtype, location, and volume of disease, it was suggested that PET‐CT and ceCT and could be used for staging, and then response assessed only with PET‐CT if avid at baseline. A similar consideration was made for staging of lymphoplasmacytic lymphoma/WM (Waldenström macroglobulinemia) (predominately nodal or skeletal presentation). In addition, PET‐CT should also be considered to detect nodal and visceral disease and to direct biopsy in cutaneous lymphomas.

3.3. Response Assessment

Visual assessment should be accompanied by semi‐quantitative assessment in the DS. There was agreement that the current definition of complete metabolic response (CMR), using DS1‐3 be retained for response assessment. It was proposed that the definition of DS5 should be at least 2 times the maximum standardized uptake value (SUV) of the liver and should be distinguished into DS5a with persistent or progressive metabolic disease (PMD) due to an increase in intensity and/or extent of existing lesions only, and DS5b with PMD due to the presence of new lesions to allow comparisons of their clinical relevance. Clarifications regarding the DS and interpretation of metabolic response categories were discussed and summarized in Table 1. There was limited enthusiasm to replace a PET‐CT based system including bidimensional lymph node measurements when needed [6] with unidimensional CT‐based system [11] which has not shown superior results [12].

TABLE 1.

Summary of suggested clarifications to the deauville score (DS) and metabolic response categories.

Response	Change from baseline/previous scan
Complete metabolic response (CMR)	DS 1–3 (DS 1 usually assigned where lesion is no longer visible on CT)
Partial metabolic response (PMR)	DS 4 or 5a with responding disease meaning reduced intensity a with no increase in extent b of lymphoma or no significant reduction in intensity but reduction in the extent of lymphoma
Stable metabolic disease (SMD)	DS 4 or 5a and intensity and extent of lymphoma stable
Progressive metabolic disease (PMD)	DS 5a with increased intensity a and/or increased extent of lymphoma
	DS 5b new lesions due to lymphoma

^a Increase in uptake in a single lesion constitutes PMD, even if there has been a response in lesions elsewhere (sometimes referred to as “mixed metabolic response”).

^b Refers to number of lesions and/or areas of uptake.

3.4. Interim Assessment

The use of interim scanning engendered considerable controversy. Participants were divided on whether to use PET‐CT for routine interim assessment of prognosis, except when progression is suspected, or when risk‐adapted therapy is indicated. If an interim scan is considered, PET‐CT is the preferred modality for avid tumors.

Whereas earlier studies in Hodgkin lymphoma demonstrated the usefulness of interim scanning, in more recent studies where interim imaging was not performed, outstanding outcomes were achieved [13]. In NHL, reacting to interim scanning has not improved outcomes [14], although in limited disease, it may allow for treatment deescalation.

3.5. Surveillance Monitoring

The consensus opinion was that routine surveillance imaging should be abandoned and imaging performed only if prompted by a suspicion or recurrence. It was proposed that, in clinical trials, imaging be performed no more frequently than 12‐weekly, except where there is a concern regarding treatment toxicity or clinical deterioration. Continuation of treatment beyond radiological progression can be considered for patients treated with immunotherapy (e.g., checkpoint inhibitors) deriving clinical benefit, despite progression of individual lesions (LyRIC approach [15]). Either biopsy, if possible, or repeat imaging at about 12 weeks for patients who continue treatment with radiological progression in the absence of clinical deterioration is suggested. The potential use of ctDNA in this context was suggested [16].

3.6. MTV and ctDNA

Radiomics, including MTV, have the potential to enhance the prognostic value of PET‐CT [10]. Published methods to standardize the measurement of MTV using a SUV of 4.0 for tumor segmentation were endorsed [17]. Adoption of MTV in clinical reporting and exploration of other radiomic markers such as disease dissemination [18] within clinical trials was strongly encouraged.

ctDNA is anticipated to play an increasing role in prognosis, surveillance and monitoring and collection at baseline, during response assessment and surveillance should be strongly encouraged in clinical trials. Current data suggest that ctDNA is complimentary to imaging [19] but might, in the future, replace imaging in certain situations, such as disease monitoring.

4. Conclusions

The Workshop and subsequent questionnaire provide the foundation for the next iteration of The Lugano Classification. While it was clear that minor modifications could be made now, it was concluded that the current classification should be retained until the standardization of techniques for measuring MTV ctDNA are firmly established and their practical contribution demonstrated. Meantime, further discussion should continue about the structure of the classification, the interface with pediatrics and chronic lymphocytic leukemia, and the best approach to primary extranodal lymphoma.

It is recommended that the ICML sponsored standing committee monitor progress and ensure comprehensive revision of the Lugano Classification when appropriate.

Conflicts of Interest

The author declares no conflicts of interest.

Peer Review

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1002/hon.70103.

Appendix 1. Questionnaire and Summary of Responses

	Statement	Total N	Strongly agree N (%)	Agree N (%)	Neutral N (%)	Disagree N (%)	Strongly disagree N (%)	Not qualified N
1	We need a revision of the lugano classification now	69	21 (30)	32 (46)	11 (16)	5 (7)	0 (0)	0
2	A revised lugano classification should also include specific recommendations about staging and restaging of extranodal lymphomas (for the most common sites)	70	22 (32)	35 (51)	5 (7)	5 (7)	2 (3)	1
3	The same staging classification should be used for all nodal lymphomas	69	22 (32)	35 (51)	4 (6)	5 (7)	2 (3)	1
4	We should maintain an anatomical 4‐stage classification	70	3 (4)	16 (23)	13 (19)	22 (32)	15 (22)	1
5	The anatomical 4‐stage classification should be replaced by a 2‐stage risk‐based one (e.g., limited stage: I‐II with low‐risk: advanced stage—stage I‐II with high risk, stages III and IV)	70	3 (4)	15 (21)	13 (19)	22 (31)	16 (23)	1
6	PET‐CT is recommended for staging and response assessment of aggressive lymphomas, including systemic T‐cell lymphoma, follicular lymphoma, and mantle cell lymphoma	68	52 (79)	14 (21)	0	0	0	2
7	Use of contrast‐enhanced CT (ceCT) is not necessary for staging and response assessment in FDG‐avid lymphoma and should be discouraged in treatment guidelines and in clinical practice, except when required for radiotherapy planning at baseline or for other specific clinical indications (e.g., assessment of compression/thrombosis of central/mediastinal vessels)	67	24 (37)	20 (31)	7 (11)	11 (17)	3 (5)	2
8	ceCT and/or PET‐CT is recommended for staging of marginal zone lymphoma (nodal, splenic, and extranodal) and PET‐CT for response assessment if FDG‐avid on baseline imaging	66	22 (35)	29 (46)	5 (8)	7 (11)	0	3
9	ceCT and/or PET‐CT is recommended for staging of lymphoplasmacytic lymphoma/Waldenstrom macroglobulinaemia (predominant nodal or skeletal presentation) and PET‐CT for response assessment if FDG‐avid on baseline imaging	66	14 (22)	28 (44)	11 (17)	9 (14)	1 (2)	3
10	PET‐CT is recommended to detect nodal and visceral involvement in cutaneous lymphomas with advanced disease and to direct biopsy	65	23 (35)	32 (49)	8 (12)	2 (3)	0	5
11	We should abandon routine interim PET (except in trials) in situations where PET‐CT provides prognostic information only, until there is evidence that altering therapy on the basis of the PET findings leads to improved patient outcome	69	10 (15)	16 (24)	11 (16)	17 (25)	14 (21)	1
12	When “interim” imaging is required, PET‐CT is recommended as the best imaging modality to assess early response	69	33 (49)	23 (34)	7 (10)	5 (7)	0	1
13	We should abandon routine surveillance imaging in CMR/CR unless there is clinical suspicion of relapse	65	26 (41)	13 (21)	7 (11)	12 (19)	5 (8)	2
14	ctDNA is anticipated to play an increasing role in surveillance monitoring and should replace routine imaging	61	11 (41)	25 (21)	31 (11)	25 (19)	8 (8)	2
15	For clinically suspected relapse, FDG PET‐CT is recommended for all indications that support FDG PET‐CT scans at baseline	70	44 (64)	24 (35)	0	1 (1)	0	1
16	The SUV4 threshold is recommended for measurement of metabolic tumor volume (MTV), due to simplicity and generalizability in aggressive subtypes and FL	69	11 (20)	25 (46)	14 (26)	4 (7)	0	15
17	Baseline MTV estimate using the benchmark method of measurement (SUV4 with minimum volume 3 mLs) should be encouraged for routine clinical reporting	69	12 (21)	28 (48)	11 (19)	7 (12)	0	11
18	The current definition of complete metabolic response (DS 1–3) should be maintained	69	16 (24)	46 (69)	2 (3)	3 (4)	0	2
19	Deauville score 5 should be defined as uptake that is at least 2 times the maximum SUV in the liver	67	9 (15)	30 (48)	20 (32)	3 (5)	0	5
20	Deauville score 5 should be subdivided into:	69	17 (25)	43 (64)	6 (9)	1 (1)	0	2
	DS5a with PMD due to an increase in intensity and/or extent of existing lesions only
	DS5b with PMD due to the presence of new lesions (to allow the clinical relevance of PMD due to increased uptake in pre‐existing lesions to be compared with that due to new lesions)
21	Response assessment should be limited to 3 categories: CMR versus no or incomplete response (NICR) versus PMD	63	3 (5)	28 (45)	12 (19)	15 (24)	4 (6)	1
	For non‐CMR, the change in SUV on PET‐CT could be recorded to inform how the delta SUV is related to response in specific clinical scenarios
22	Response assessment should include 4 categories (CMR, PMR, SMD, PMD) to differentiate responding from non‐responding disease without clear progression and the table can be used to clarify how response categories might be assigned a	65	14 (22)	35 (55)	11 (17)	4 (6)	0	1
23	Use of contrast enhanced CT where PET‐CT is the primary method of response assessment should be discouraged in clinical trials, except where required for radiotherapy planning at baseline	70	27 (39)	31 (45)	5 (7)	5 (7)	1 (6)	1
24	Eligibility for inclusion in clinical trials can be based on the presence of FDG‐avid assessable disease attributable to lymphoma. Contrast enhanced CT is not required where there is evaluable disease on PET‐CT	70	32 (46)	29 (42)	5 (7)	2 (3)	1 (1)	1
25	Bidimensional measurements can be replaced by unidimensional measurements in non‐FDG avid lymphoma, where ceCT is the primary method of assessment	70	9 (13)	30 (44)	18 (26)	8 (12)	3 (4)	2
26	The conventional use of a maximum of 6 target measurable lesions should be kept in non‐FDG avid lymphoma	70	5 (7)	28 (41)	22 (32)	13 (19)	0	2
27	Continuation of treatment beyond radiological progression should be allowed for patients treated with immunotherapy (e.g., checkpoint inhibitors) deriving clinical benefit, where there is stability in overall tumor burden, despite progression of individual lesions (LyRIC approach)	69	15 (22)	43 (64)	7 (10)	2 (3)	0	2
28	MTV in FDG‐avid aggressive lymphoma rather than ceCT could be considered to assess tumor burden with checkpoint inhibitors and other immune therapies	69	9 (14)	37 (56)	18 (27)	2 (3)	0	3
29	To allow comparison and pooling of clinical trial results, SUV4 is the agreed benchmark standard for segmentation of TMTV (and other metrics such as TLG and Dmax), and measurement of these metrics should be systematically reported	67	9 (16)	33 (58)	14 (25)	1 (2)	0	10
30	Repeat imaging at 12 weeks for all patients who continue treatment with radiological progression in the absence of clinical deterioration is suggested in clinical trials to establish a reliable cut‐off for prediction of true progression using MTV	67	11 (17)	43 (65)	9 (14)	2 (3)	1 (2)	1
31	Imaging should be performed no more frequently than 12‐weekly, except where there is a concern regarding treatment toxicity or clinical deterioration	68	26 (39)	36 (54)	2(3)	3 (4)	0	1
32	ctDNA collection at baseline and during each response assessment imaging should be strongly encouraged in clinical trials	70	31 (44)	33 (47)	5 (7)	1 (1)	0	0

Note: Some percentages may not add up to 100% due to rounding errors. Italics values (by brakets) are indicating percentages in the subsequents rows.

Abbreviations: ctDNA, circulating tumor DNA; CT, computed tomography; D_max, maximum distance between lesions in the body; DS, Deauville score; FDG, 18fluorine‐fluorodeoxyglucose; LyRIC, lymphoma response to immunomodulatory thearpy criteria; N, total number of respondents; PET, positron emission tomography; SUV, standardized uptake value; TMTV, total metabolic tumor volume; %, percentage of respondents expressing an opinion (excluding those who responded “not qualified”).

^a see Table 1.

Appendix 2. Subcommittee Members

Nodal Subcommittee

B.D. Cheson (Chair)

R. Advani

S. Ansell

P. Armand

M. Dreyling

J. Friedberg

S. Horwitz

R. Hoppe

P.W.M. Johnson

A. LaCasce

J. Radford

L. Sehn

Extranodal Lymphoma Subcommittee

E. Zucca (Chair)

A. Ferreri

T. Habermann

I. Lossos

M. Raderer

S. Smith

C. Thieblemont

A. Zelenetz

Imaging Subcommittee

S. Barrington (Chair)

P. Armand

R. Boellaard

R.O. Casasnovas

L. Ceriani

A‐S. Cottereau

A. Ferrajoli

L. Guerra

M. Hutchings

C. Kobe

L. Kostakoglu‐Shields

G. Mikhaeel

L. Schwartz

ctDNA/Genetics/MRD Subcommittee

D. Rossi (Chair)

A. Alizadeh

E. Campo

P. Ghia

M. Ladetto

C. Pott

M. Roschewski

D. Scott

M. Shipp

Pediatric Subcommittee

K. Kelly (Chair)

J. Flerlage

B. Hoppe

R. Kluge

C. Mauz‐Koerholz

W. Wössmann

Appendix 3. Participants at the ICML Workshop

Advani Ranjana	Stanford	US
Alizadeh Ash A.	Stanford	US
Ansell Steve	Rochester	US
Armand Philippe	Boston	US
Barrington Sally	London	GB
Boellaard Ronald	Amsterdam	NL
Borchmann Peter	Cologne	DE
Campo Elias	Barcelona	ES
Casasnovas Olivier	Dijon	FR
Ceriani Luca	Bellinzona	CH
Cheson Bruce D.	Rockville	US
Cottereau Anne Segolene	Paris	FR
Cwynarski Kate	London	GB
Davies Andrew	Southampton	GB
Dreyling Martin	Munich	DE
Evens Andrew	Boston	US
Ferrajoli Alessandra	Houston	US
Ferreri Andrés	Milan	IT
Flerlage Jaime	Memphis	US
Fowler Nathan	Houston	US
Friedberg Jonathan W.	Rochester	US
Gallamini Andrea	Nice	FR
Ghia Paolo	Milan	IT
Guerra Luca	Monza	IT
Habermann Tom	Rochester	US
Herrera Alex	Duarte	US
Hoppe Richard	Los Altos	US
Hutchings Martin	Copenhagen	DK
Johnson Peter W.M.	Southampton	GB
Kelly Kara	Buffalo	US
Kersten Marie José	Amsterdam	NL
Kim Won Seog	Seoul	KR
Kimby Eva	Östhammar	SE
Kluge Regine	Leipzig	DE
Kobe Carsten	Cologne	DE
Kostakoglu Shields Lale HS	New York	US
LaCasce Ann	Boston	US
Ladetto Marco	Alessandria	IT
Leonard John P.	New York	US
Lister T. Andrew	London	GB
Lopez‐Guillermo Armando	Barcelona	ES
Lossos Izidore S.	Miami	US
Luminari Stefano	Reggio Emilia	IT
Lunning Matthew	Omaha	US
Maruyama Dai	Tokyo	JP
Mauz‐Koerholz Christine	Gießen	DE
Mikhaeel George	London	GB
Montoto Silvia	London	GB
Morschhauser Frank	Lille	FR
Moskowitz Alison	New York	US
Moskowitz Craig	Miami	US
Nastoupil Loretta	Houston	US
Pavlovsky Astrid	Buenos Aires	AR
Pott Christiane	Kiel	DE
Radford John	Manchester	GB
Roschewski Mark	Bethesda	US
Rossi Davide	Bellinzona	CH
Salles Gilles	New York	US
Savage Kerry	Vancouver	CA
Schwartz Lawrence	New York	US
Scott David	Vancouver	CA
Sehn Laurie	Vancouver	CA
Seymour John F.	Melbourne	AU
Shipp Margaret	Boston	US
Smith Sonali	Chicago	US
Thieblemont Catherine	Paris	FR
Trneny Marek	Prague	CZ
Trotman Judith	Concord	AU
Vose Julie	Omaha	US
Wössmann Wilhelm	Hamburg	DE
Zelenetz Andy	New York	US
Zhao Weili	Shanghai	CN
Zinzani Pier Luigi	Bologna	IT
Zucca Emanuele	Bellinzona	CH

Data Availability Statement

The data supporting the findings of this manuscript are available from the corresponding author upon reasonable request.