Ketone Bodies Are Potential Prognostic Biomarkers in Relapsed/Refractory Diffuse Large B-Cell Lymphoma: Results from the R2-GDP-GOTEL Trial

Academic Edi o : Alexanda Tzanko Recei ed: 3 Decembe 2024 Re ised: 10 Janua y 2025 Accep ed: 16 Janua y 2025 Published: 5 Feb ua y 2025 Ci a ion: Fe nández-Cas illejo, S.; Badia, J.; de la C uz-Me ino, L.; Ma ín Ga cia-Sáncho, A.; Ca nice o- González, F.; Palazón-Ca ión, N.; Ríos-He anz, E.; de la C uz-Vicen e, F.; Rueda-Domínguez, A.; Ma ínez- Banaclocha, N.; e al. Ke one Bodies A e Po en ial P ognos ic Bioma ke s in Relapsed/Re ac o y Di use La ge B-Cell Lymphoma: Resul s om he R2-GDP-GOTEL T ial. Cance s 2025, 17, 532. h ps://doi.o g/10.3390/ cance s17030532 Copy igh : © 2025 by he au ho s. Licensee MDPI, Basel, Swi ze land. This a icle is an open access a icle dis ibu ed unde he e ms and condi ions o he C ea i e Commons A ibu ion (CC BY) license (h ps://c ea i ecommons.o g/ licenses/by/4.0/). A icle Ke one Bodies A e Po en ial P ognos ic Bioma ke s in Relapsed/ Re ac o y Di use La ge B-Cell Lymphoma: Resul s om he R2-GDP-GOTEL T ial Sa a Fe nández-Cas illejo 1,2,† , Joan Badia 1,2,† , Luís de la C uz-Me ino 3,4 , Alejand o Ma ín Ga cia-Sáncho 5,6 , Fe nando Ca nice o-González 7 , Na alia Palazón-Ca ión 3,4 , Edua do Ríos-He anz 8 , Fá ima de la C uz-Vicen e 9 , An onio Rueda-Domínguez 10 , Na i idad Ma ínez-Banaclocha 11 , José Gómez-Codina 12, Jo ge Lab ado 13 , F ancisca Ma ínez-Madueño 1,2,14, Nú ia Amigó 14,15 , An onio Sala -Sil es e 16, Del ys Rod íguez-Ab eu 17, Lau a Gál ez-Ca ajal 10 , Ma ga i a Sánchez-Bea o 18 , Ma iano P o encio-Pulla 19 , Ma ia Gui ado-Risueño 20 , Es eban Nogales 3,4, Víc o Sánchez-Ma gale 21 , Ca los Jiménez-Co egana 21 , Guille mo Rod íguez-Ga cía 9, Raquel Cume as 1,22,* and Josep Gumà1,2,14 on behal o he Spanish Lymphoma Oncology G oup (GOTEL) 1 T ansla ional, Epidemiological and Clinical Oncological Resea ch G oup (GIOTEC), Depa men o Oncology, Ins i u d’In es igació Sani à ia Pe e Vi gili (IISPV), 43204 Reus, Ta agona, Spain; sa a. e nandez@iisp .ca (S.F.-C.); joan.badia@iisp .ca (J.B.); [email p o ec ed] (F.M.-M.); [email p o ec ed] (J.G.) 2Ins i u d’Oncologia de la Ca alunya Sud (IOCS), Hospi al Uni e si a i San Joan de Reus, 43204 Reus, Ta agona, Spain 3Cance Immuno he apy G oup, Oncohema ology and Gene ics Depa men , Biomedicine Ins i u e o Se ille (IBIS)/CSIC, 41013 Se ille, Spain; [email p o ec ed] (L.d.l.C.-M.); [email p o ec ed] (N.P.-C.); [email p o ec ed] (E.N.) 4 Depa men o Clinical Oncology, Uni e si y Hospi al Vi gen Maca ena and School o Medicine, Uni e si y o Se illa, 41013 Se illa, Spain 5Depa men o Hema ology, Hospi al Uni e si a io de Salamanca, Ins i u o de In es igación Biomédica de Salamanca (IBSAL), Uni e sidad de Salamanca, 37007 Salamanca, Spain; [email p o ec ed] 6CIBER de Cánce (CIBERONC), Ins i u e o Heal h Ca los III, 28029 Mad id, Spain 7Depa men o Hema ology, San Ped o de Alcán a a Hospi al, 10003 Cáce es, Spain; ca nice [email p o ec ed] 8Depa men o Hema ology, Hospi al Uni e si a io Vi gen de Valme, 41014 Se illa, Spain; [email p o ec ed] 9Depa men o Hema ology, Hospi al Uni e si a io Vi gen del Rocío, 41013 Se illa, Spain; [email p o ec ed] (F.d.l.C.-V.); [email p o ec ed] (G.R.-G.) 10 Depa men o Clinical Oncology. Hospi al Uni e si a io Vi gen de la Vic o ia, 29010 Málaga, Spain; [email p o ec ed] (A.R.-D.); [email p o ec ed] (L.G.-C.) 11 Depa men o Oncology, D . Balmis Gene al Uni e si y Hospi al, Alican e Ins i u e o Heal h and Biomedical Resea ch (ISABIAL), 03010 Alican e, Spain; [email p o ec ed] 12 Depa men o Clinical Oncology, Hospi al Uni e si a io y Poli écnico La Fe, 46026 Valencia, Spain; [email p o ec ed] 13 Depa men o Hema ology, Hospi al Uni e si a io de Bu gos, 09006 Bu gos, Spain; [email p o ec ed] 14 Facul y o Medicine and Heal h Sciences, Uni e si a Ro i a i Vi gili (URV), 43201 Reus, Ta agona, Spain; [email p o ec ed] 15 Bios e Teslab, 43206 Reus, Ta agona, Spain 16 Depa men o Hema ology, Hospi al del Ma , 08003 Ba celona, Spain; asala @hospi aldelma .ca 17 Depa men o Clinical Oncology, Hospi al Uni e si a io Insula de G an Cana ia, 35016 Las Palmas de G an Cana ia, Las Palmas, Spain; [email p o ec ed]g 18 Lymphoma Resea ch G oup, Depa men o Medical Oncology, Hospi al Uni e si a io Pue a de Hie o-Majadahonda, IDIPHISA, 28222 Majadahonda, Mad id, Spain; [email p o ec ed]g 19 Depa men o Clinical Oncology, Hospi al Uni e si a io Pue a De Hie o-Majadahonda, IDIPHISA, 28222 Majadahonda, Mad id, Spain; [email p o ec ed] 20 Depa men o Clinical Oncology, Hospi al Uni e si a io de Elche, 03203 Elche, Alican e, Spain; [email p o ec ed] 21 Medical Biochemis y and Molecula Biology and Immunology, Hospi al Uni e si a io Vi gen de la Maca ena, 41009 Se illa, Spain; [email p o ec ed] (V.S.-M.); [email p o ec ed] (C.J.-C.) 22 Depa men o Elec ical and Au oma ic Elec onic Enginee ing, Uni e si a Ro i a i Vi gili (URV), 43002 Ta agona, Spain *Co espondence: aquel.cume as@iisp .ca †These au ho s con ibu ed equally o his wo k. Cance s 2025,17, 532 h ps://doi.o g/10.3390/cance s17030532 Cance s 2025,17, 532 2 o 16 Simple Summa y: Pa ien s wi h elapsed o e ac o y di use la ge B-cell lymphoma (DLBCL) ha e poo ou comes and limi ed ea men op ions. A phase II ial by GOTEL e alua ed he R2-GDP egimen (combina ion o lenalidomide, i uximab, gemci abine, dexame hasone, and cispla in), demons a ing easibili y and e ec i eness. Baseline se um me abolomic analysis o 69 pa ien s en olled in he ial iden i ied wo independen me abo- li es, 3-hyd oxybu y a e (3OHB) and ace one, as being signi ican ly associa ed wi h o e all su i al and p og ession- ee su i al. Ele a ed 3OHB le els (>141 µ M) we e speci ic o he ABC sub ype o DLBCL, while ace one le els we e ele a ed in bo h ypes o DLCBL bu mo e p onounced in ABC cases. These bioma ke s, i espec i e o sex, age, and BMI, could help p edic ou comes and guide ea men s a egies in elapsed/ e ac o y DLBCL. Abs ac : Backg ound: Pa ien s wi h elapsed o e ac o y (R/R) di use la ge B-cell lymphoma (DLBCL) who a e ineligible o high-dose chemo he apy ha e limi ed ea - men op ions and poo li e expec ancy. The pu pose o his s udy is o iden i y a se um me abolomic p o ile ha may be p edic i e o ou come in pa ien s wi h R/R-DLBCL. Me hods: This s udy included 69 R/R DLBCL pa ien s om he R2-GDP-GOTEL ial (Eud aCT 2014-001620-299). Se um samples we e collec ed a baseline, and he mean leng h o ollow-up was 41 mon hs. Se um me aboli es we e analyzed by nuclea magne ic esonance (NMR). Me aboli es we e co ela ed wi h ea men esponse, p og ession- ee su i al (PFS), and o e all su i al (OS). Resul s: Se um le els o 3-hyd oxybu y a e (3OHB) and ace one we e signi ican ly (p< 0.001) associa ed wi h PFS (3OHB: haza d a io [HR] 7.7, 95% con idence in e al [CI] 2.5–24.1; ace one: HR 9.32, 95% CI 2.75–31.6) and OS (3OHB: HR 9.32, 95% CI 2.75–31.6; ace one: HR 1.92, 95% CI 1.36–2.69). Se um alues o 141 µ M o 3OHB and 40 µ M o ace one we e he op imal cu o s associa ed wi h he su i al ou comes. Ele a ed 3OHB le els (>141 µ M) we e speci ic o he ABC sub ype o DLBCL, while ace one le els we e ele a ed in bo h ypes o DLCBL bu mo e p onounced in ABC cases. In a mul i a ia e su i al analysis, including he In e na ional P ognos ic Index (IPI) sco e and e ac o iness s a us (R/R), 3OHB and ace one emained signi ican . To aid oncologis s employing he R2-GDP egime, we cons uc ed PFS and OS nomog ams o R/R-DLBCL isk s a i ica ion, inco po a ing 3OHB le els o ace one le els, IPI sco e, and e ac o iness s a us. The nomog am wi h 3OHB and e ac o iness s a us showed a ime-dependen AUC o 0.86 o 6-mon h PFS and 0.84 o 12-mon h OS. These nomog ams p o ide a comp ehensi e ool o indi idualized isk assessmen and ea men op imiza ion. Conclusions: The ke one bodies 3OHB and ace one a e po en ial p ognos ic bioma ke s o poo ou come in R/R DLBCL pa ien s ea ed wi h he R2-GDP egimen, independen ly o IPI sco e and chemo e ac o iness s a us. Keywo ds: ke one bodies; di use la ge B-cell lymphoma (DLBCL); elapsed/ e ac o y lymphoma; 3-hyd oxybu y a e; 3OHB; ace one; p ognos ic bioma ke s; me abolomics 1. In oduc ion App oxima ely 60% o pa ien s wi h di use la ge B-cell lymphomas (DLBCL) a e cu ed using up on he apy wi h he CHOP-R egimen o o he an h acycline and i uximab- based chemo he apies, whe eas he emaining 40% a e e ac o y o elapsed (R/R) ol- lowing i s -line chemo he apy. A he ime his clinical ial was conduc ed, he s anda d ea men o R/R DLBCL pa ien s was second-line con en ional chemo he apy ollowed by consolida ion wi h high-dose chemo he apy in chemosensi i e pa ien s. Pallia i e chemo he apy is an op ion o pa ien s who a e unable o ecei e high-dose chemo he apy Cance s 2025,17, 532 3 o 16 o CAR-T he apy; ne e heless, he bes he apeu ic op ion may in ol e en olling he pa ien in in es iga ional clinical ials. The iden i ica ion o bioma ke s capable o p edic ing he ou come o DLBCL pa ien s has been he ocus o inc easing in e es due o he ma ked gene ic and molecula he e o- genei y ha unde lies disease agg essi eness and umo p og ession. Me abolomics is a powe ul ool ha can iden i y cance bioma ke s and d i e s o umo igenesis. In he ield o lymphomas, di e en s udies ha e e alua ed un a ge ed me abolomics using gas (GC) o liquid ch oma og aphy (LC) coupled wi h mass spec ome y (MS) in pa ien s wi h lymphoid neoplasms and heal hy popula ions [ 1 – 5 ]. Al hough mos s udies showed me abolomic di e ences be ween pa ien s and heal hy subjec s, he iden i ica ion o di - e en ial me aboli es has been inconsis en . This is likely due o di e ences in labo a o y echniques (GC-MS and LC-MS) and/o biological samples (blood, u ine, and eces). How- e e , he me abolomic p o ile in pa ien s wi h DLBCL as a p ognos ic ac o o su i al has been e alua ed in only wo s udies. S enson e al. [ 6 ] used nuclea magne ic esonance (NMR) spec oscopy in 87 pa ien s wi h DLBCL p io o s a ing i s -line ea men wi h chemoimmuno he apy. S a is ically signi ican di e ences we e ound in he me abolomic p o ile be ween pa ien s who achie ed a comple e esponse and long su i al and hose who we e e ac o y o ea men o elapsed. Pa ien s who had been cu ed showed highe le els o aspa a e, aline, o ni hine, and py oglu ama e, whe eas R/R pa ien s had highe concen a ions o lysine, a ginine, cada e ine, and 2-hyd oxybu y a e. In a second s udy, Mi e al. [ 7 ] used GC-MS o assess p e- ea men se um samples om 80 DLBCL pa ien s and epo ed ha highe le els o py oglu amic and hexadecenoic acids and lowe le els o aline we e associa ed wi h signi ican ly highe o e all su i al. The Spanish G oup o he T ea men and S udy o Lymphomas (GOTEL) conduc ed a phase II clinical ial o e alua e he combina ion o lenalidomide wi h R-GDP ( i uximab, gemci abine, dexame hasone, and cispla in) in pa ien s wi h R/R DLBCL who we e ei he unsui able o high-dose chemo he apy o whose ea men had no wo ked. A o al o 78 pa ien s we e included in he R2-GDP-GOTEL s udy, and a e a median ollow-up o 37 mon hs, 7.9% o pa ien s we e s ill ali e wi hou p og ession a 24 mon hs [ 8 ]. Taking ad an age o baseline da a om his clinical ial, he p esen s udy was designed o iden i y a se um me abolomic p o ile ha may be p edic i e o ou comes in pa ien s wi h R/R DLBCL ea ed wi h he R2-GDP combina ion. 2. Ma e ials and Me hods 2.1. S udy Design and Pa ien s The R2-GDP-GOTEL clinical ial was a phase II, mul icen e , open-label, and single- a m s udy ca ied ou in 78 R/R DLBCL pa ien s who we e ea ed in 18 Spanish hospi als be ween Ap il 2015 and Sep embe 2018. B ie ly, he R2-GDP egimen included an induc ion ea men wi h a combina ion o lenalidomide, i uximab, gemci abine, dexame hasone, and cispla in (R2-GDP), o up o 6 cycles (e e y 3 weeks), ollowed by main enance wi h lenalidomide o up o 24 mon hs, unless he e was a p og ession, unaccep able oxici y, o olun a y wi hd awal [ 8 ]. Addi ional in o ma ion abou he d opou a e (a i ion a e) and a powe analysis o he R2-GDP ial ha e been p e iously epo ed [ 8 ]. Since his ial has a single-a m design, andomiza ion and blinding do no apply. Eligible pa icipan s we e pa ien s wi h chemo e ac o y o elapsed DLBCL unsui able o high-dose chemo he apy, wi h an Eas e n Coope a i e Oncology G oup (ECOG) pe o mance s a us o 0–1, and who had p e iously ecei ed a leas one line o immunochemo he apy, including i uximab. Pa ien s wi h lep omeningeal o cen al ne ous sys em (CNS) in ol emen , and hose wi h hema ological, enal, o li e dys unc ion, we e excluded om he ini ial R2-GDP-GOTEL clinical ial. Since all he lymphomas included in his s udy we e DLBCL, all he included Cance s 2025,17, 532 4 o 16 pa ien s exp essed CD45 and he pan B-cell ma ke s CD19, CD20, CD22, and CD79a in he ini ial diagnos ic immunohis ochemical s udy. The objec i e o he p esen subs udy was o assess he baseline me abolomic p o ile o R/R DLBCL pa ien s and o iden i y se um me aboli es p edic i e o ou come. Fo ha pu pose, 69 pa ien s om he R2-GDP-GOTEL clinical ial o whom a su icien blood sample was a ailable o me abolomic p o iling we e included in he me abolomic p o iling subs udy. 2.2. Sample P epa a ion and Me abolomic P o iling A comp ehensi e me hodological app oach o high- h oughpu sc eening by NMR spec oscopy using a B uke A ance 600 MHz NMR spec ome e (B uke BioSpin, E lingen, Ge many ) was used o analyze a b oad spec um o me aboli es in se um sam- ples [ 9 ]. The analysis included he lipop o ein, glycop o ein, and me aboli e p o iles om in ac se um, in addi ion o he lipid p o ile om lipid se um ex ac s ob ained by a biphasic ex ac ion wi h me hyl e -bu yl e he (MTBE). Lipid ex ac s we e d ied and econs i u ed in 0.01% e ame hylsilane (TMS) solu ion (0.067 mM) and deu e a ed sol en s. All analy- ses we e ca ied ou a Bios e Teslab (Reus, Ta agona, Spain). In addi ion, 1D Nuclea O e hause E ec Spec oscopy (NOESY) was used o cha ac e ize small molecules such as amino acids and small ca bohyd a es, while la ge molecules like lipop o eins and glyco- p o eins we e de ec ed using LED Di usion (Di ) expe imen s. All he sequences an a 37 ◦C in quan i a i e condi ions. Samples we e coded o main ain he subjec s’ anonymi y. The Liposcale ® es (IVD-CE) was used o de e mine he lipid composi ion, pa icle size, and concen a ion o majo lipop o ein classes as well as he pa icle concen a ion o nine subclasses [ 10 ]. Ci cula ing glycop o eins we e ob ained by decon olu ing he speci ic NMR spec al egions and quan i ying a eas co ela ing o he concen a ion o he ace yl g oups o N-ace ylglucosamine and N-ace yl galac osamine (GlycA) and ace yl g oups o N-ace ylneu aminic acid (GlycB). A Ca –Pu cell–Meiboom–Gill (CPMG) il e on he 1H-NMR spec a was used o p o ile and absolu ely quan i y he me abolomics p o ile. The BUME p o ocol [ 11 ] was used o lipid quan i ica ion, based on he Lipspin so wa e [ 12 ]. Succinc ly, we used lineshape i ing analysis o spec al egions o quan i y he lipids. 2.3. Endpoin s The p ima y endpoin s we e he cha ac e iza ion o he baseline me abolomic p o ile in R/R DLBCL pa ien s acco ding o esponse o ea men and ou come. Key seconda y endpoin s we e o de e mine he p edic i e pe o mance o nomog ams o R/R DLBCL isk s a i ica ion based on he me aboli es iden i ied and hei op imal cu o s associa ed wi h he ou come. Tumo esponse was e alua ed acco ding o he In e na ional Wo king G oup C i e ia [ 13 ] using compu ed omog aphy (CT) a e he hi d induc ion cycle and posi on emission omog aphy (PET) in he ollowing 4 weeks a e he las cycle o he induc ion phase. Ou come included p og ession- ee su i al (PFS) and o e all su i al (OS). PFS was de ined as he ime be ween he i s dose o he R2-GDP schedule o he p og ession o disease o dea h, and OS was de ined as he pe iod om he i s dose o he R2-GDP schedule o dea h om any cause. 2.4. S a is ical Analysis S a is ical analyses we e pe o med wi h R so wa e using he R S a s Package ( .4.3.2.), su i al ( .3.5-7), su mine ( .0.4.9), maxs a ( .0.7-25), ca e ( .6.0-94), ms ( .6.7-1), and su i alROC ( .1.0.3.1). The analyses we e es ic ed o me aboli es iden i ied in >90% o pa ien s. Me aboli e missing alues we e no impu ed, and me aboli e concen a ions we e only scaled in mul i a ia e analyses. Fo ep oducibili y, he andom seed was se a 123 ia he se .seed unc ion (R base). Uni a ia e s a is ics was pe o med using he non-pa ame ic Cance s 2025,17, 532 5 o 16 Mann–Whi ney–Wilcoxon s a is ical es . Su i al analysis was ca ied ou using he non- pa ame ic Kaplan–Meie me hod and he log- ank es o he compa ison o su i al cu es. Cu o s o nume ical a iables we e calcula ed using he maximally selec ed ank s a is ics, wi h a 10% minimum p opo ion o obse a ions, and missing alues we e assigned o he lowes g oup. Cox eg ession analyses we e un wice: once o sea ch o s a is ically signi ican indi idual p ognos ic me aboli es, and again o con i m whe he hey emained s a is ically signi ican in a mul i a ia e analysis wi h o he s a is ically signi ican clinical p ognos ic ac o s, such as he In e na ional P ognos ic Index (IPI) [ 14 ] and e ac o iness s a us. The haza d a io (HR) and he 95% con idence in e al (CI) we e calcula ed. S a is ical signi icance was se a p< 0.05, and a alse disco e y a e (FDR) p alue adjus men was applied when necessa y. Fold change (FC) analyses we e used o compa e he absolu e alue o change in he means o each me aboli e be ween esponde s and non- esponde s. The nomog am o R/R DLBCL isk s a i ica ion inco po a ed he signi ican me abo- li es iden i ied in he uni a ia e Cox analysis and he signi ican clinical isk ac o s wi hin a Cox p opo ional haza ds amewo k. A 70% ain spli was used o model i ing ia boo s ap calib a ion (B = 1000), while he emaining 30% es spli was used o p edic i e pe o mance e alua ion wi h a ime-dependen ecei e ope a ing cha ac e is ic cu e (ROC) and a ea unde he ROC (AUC). The ime poin s selec ed we e 6 mon hs o PFS and 12 mon hs o OS. Pea son’s p oduc -momen co ela ion coe icien ( ) was used o assess he ela ionship be ween he me aboli es iden i ied. 3. Resul s 3.1. Baseline Cha ac e is ics o Pa ien s and Su i al Six y-nine pa icipan s in he R2-GDP-GOTEL clinical ial (35 men and 34 women) wi h a median age o 70 yea s we e included in he me abolomic p o iling s udy. Key base- line da a a e displayed in Table 1. Ac i a ed B-cell-like (ABC) lymphomas, elapsed DLBCL, and IPI low/high-in e media e isk ca ego y (0–3) we e he mos common cha ac e is ics. Table 1. Key cha ac e is ics o he s udy popula ion. Va iables 1To al Pa ien s (n= 69) Sex, n(%) Men 35 (50.7%) Women 34 (49.3%) Age, yea s, median (IQR) 70.1 (61.7–74.5) Body mass index (BMI), kg/m2, median (IQR) 27.0 (23.9–30.2) Cell-o -o igin (CoO), n(%) [n= 65] Ge minal cen e B-cell-like 27 (41%) Ac i a ed B-cell-like 39 (59%) Re ac o iness, n(%) Relapsed 40 (58.0%) Chemo e ac o y 29 (42.0%) In e na ional P ognos ic Index (IPI) sco e, n(%) Low isk (0–1) 12 (17.4%) Low/High-in e media e isk (2–3) 37 (53.6%) High isk (4–5) 20 (29.0%) Response, n(%) Responde s 41 (59.4%) Non- esponde s 28 (40.6%) 1IQR: in e qua ile ange (25 h–75 h pe cen ile). In his subse o pa ien s, he o e all esponse o R2-GPD ea men was 59.4% ( n= 41 ) (comple e esponse 37.7%, pa ial esponse 21.7%), simila o hose p e iously epo ed in he whole R2-GDP popula ion [ 8 ]. The indi idual clinical e olu ion o he 69 pa ien s Cance s 2025,17, 532 6 o 16 is p esen ed in Supplemen a y Figu e S1. A e a median ollow-up o 41 mon hs, he median PFS was 5 mon hs (36%, 16%, and 7.9% a 6, 12, and 24 mon hs, espec i ely) and he median OS was 12 mon hs (66%, 47%, and 36% a 6, 12, and 24 mon hs, espec i ely) (Supplemen a y Figu e S2A,B). Su i al analyses s a i ied by cell-o -o igin (CoO) showed a median PFS o 6.0 mon hs (95% CI 3.0–11) o he ge minal cen e B-cell-like (GBC) sub ype and 5.0 mon hs (95% CI 2–6) o ABC (p= 0.099) (Figu e S2C), simila o hose p e iously epo ed in he whole R2-GDP popula ion [ 8 ]. The median OS was 16 mon hs (95% CI 6.2—no eached) o GBC and 9.3 mon hs (95% CI 5.3–24) o ABC (p= 0.59) (Figu e S2D). Conce ning DLBCL s a us, pa ien s wi h chemo e ac o y disease showed a median PFS o 3 mon hs (95% CI 2–6), whe eas hose wi h elapsed disease had a median PFS o 6 mon hs (95% CI 4–9) (p= 0.083) (Figu e S2E). The OS was signi ican ly longe o he elapsed g oup (median 24 mon hs, 95% CI 12—no eached) han o he chemo e ac o y g oup (median 6.2 mon hs, 95% CI 4.5–13) (p= 0.0034) (Figu e S2F). Pa ien s in he low/high-in e media e- isk IPI ca ego y showed a median PFS o 6 mon hs (95% CI 4–9), whils pa ien s in he high- isk ca ego y had a median PFS o 2 mon hs (95% CI 2–7) (p= 0.0044) (Figu e S2G). The median OS was also longe o he low/high-in e media e- isk IPI ca ego y (median 15 mon hs, 95% CI 9.3–33) compa ed wi h he high- isk IPI ca ego y (median 3.5 mon hs, 95% CI 1.8—no eached) (p= 0.074) (Figu e S2H). 3.2. T ea men Response Me abolomic P o ile A o al o 66 me aboli es we e iden i ied, 7 o which we e excluded as hey we e de- ec ed in less han 90% o he pa ien s. In Supplemen a y Table S1, he mean alues o he ea men esponse o he 59 included me aboli es a e p esen ed. Twel e me aboli es we e signi ican ly di e en (p< 0.05) in esponde s s. non- esponde s, including c ea inine, lac a e, glycop o ein A, ee choles e ol, es e i ied choles e ol, in e media e-densi y lipop o- ein (IDL) choles e ol, IDL iglyce ides, high-densi y lipop o ein (HDL) iglyce ides, medium and la ge low-densi y lipop o eins (LDLs), and medium and la ge HDLs. All me aboli es excep o c ea inine showed signi ican ly lowe mean alues in esponde s o R2-GPD ea men han in non- esponde s. By con as , c ea inine was signi ican ly highe in esponde s han in non- esponde s (Table 2). The box plo s o signi ican me aboli e among esponde s and non- esponde s a e shown in Supplemen a y Figu e S3. Table 2. Signi ican se um me aboli es in esponde s and non- esponde s. Me aboli es 1Responde s (n= 41) Non-Responde s (n= 28) pValue Fold Change 2 Low molecula weigh C ea inine, µM, mean (SD) 73.0 (43.3) 48.7 (17.6) 0.030 1.50 Lac a e, µM, mean (SD) 1174.2 (989.7) 2023.5 (1655.6) 0.034 0.58 Glycop o eins Glycop o ein A, µmol/L, mean (SD) 846.0 (165.3) 915.5 (161.0) 0.028 0.92 Glycop o ein A, H/W a io, mean (SD) 24.7 (4.3) 27.04 (4.7) 0.024 0.92 Choles e ol F ee choles e ol, mmol/L, mean (SD) 2.36 (0.71) 2.84 (0.83) 0.016 0.83 Es e i ied choles e ol, mmol/L, mean (SD) 4.76 (1.20) 5.36 (1.40) 0.043 0.89 Lipop o eins IDL choles e ol, mg/dL, mean (SD) 14.14 (5.84) 17.39 (5.61) 0.015 0.86 IDL iglyce ides, mg/dL, mean (SD) 13.38 (4.59) 15.64 (4.07) 0.018 0.81 LDL iglyce ides, mg/dL, mean (SD) 18.94 (6.40) 22.90 (6.08) 0.005 0.83 Medium LDL-P, nmol/L, mean (SD) 340.83 (134.98) 424.09 (157.02) 0.008 0.80 La ge LDL-P, nmol/L, mean (SD) 180.15 (44.78) 201.32 (51.39) 0.043 0.89 Medium HDL-P, µmol/L, mean (SD) 10.18 (1.68) 11.27 (1.72) 0.006 0.90 La ge HDL-P, µmol/L, mean (SD) 0.30 (0.05) 0.33 (0.05) 0.005 0.90 1 SD: s anda d de ia ion; H/W: heigh /wid h o he NMR peak; IDL: in e media e-densi y lipop o ein; LDL: low-densi y lipop o ein; HDL: high-densi y lipop o ein; P: pa icle numbe . 2 Fold change (FC) > 1 indica es an inc ease in me aboli e concen a ion and hose wi h FC < 1 indica e a dec ease in esponde s s. non- esponde s. Cance s 2025,17, 532 7 o 16 3.3. Su i al Ou come and Me abolomic P o ile Two me aboli es, 3OHB and ace one, we e signi ican ly associa ed wi h su i al ou comes in he Cox uni a ia e eg ession analysis. Highe se um concen a ions o 3OHB o ace one we e s a is ically signi ican (p< 0.001) p ognos ic ac o s o a wo se PFS (3OHB: haza d a io [HR] = 7.7; ace one: HR = 1.83) (Figu e 1A) and OS (3OHB: HR = 9.32; ace one: HR = 1.92) (Figu e 1B). Bo h 3OHB and ace one we e independen p edic o s ha we e signi ican ly associa ed wi h PFS and OS in he mul i a ia e Cox eg ession model, which included IPI isk ca ego ies and e ac o iness s a us (R/R) as classical DLBCL clinical p ognos ic ac o s (Figu e 1C–F). In he o e all s udy popula ion, he mean se um le els o 3OHB we e 120.6 µ M ( ange 0–1455.9 µ M) (n= 67) and he mean le els o ace one 40.48 µM ( ange 0–534.2 µM) (n= 68). Cance s 2025, 17, x FOR PEER REVIEW 8 o 18 Figu e 1. 3-Hyd oxybu y a e and ace one as p ognos ic me aboli es in he Cox uni a ia e analysis (A,B) and mul i a ia e eg essions models (C–F) o p og ession- ee su i al (PFS) (A,C,E) and o e all su i al (OS) (B,D,F). The op imal cu offs o se um 3OHB and ace one we e iden i ied wi h an ou come- o ien ed me hod (see Me hods) and we e se a 141 µM and 40 µM, espec i ely. Follow- ing he applica ion o he cu off, he PFS and OS Kaplan–Meie su i al cu es we e sig- ni ican ly diffe en . Fo se um 3OHB, he median PFS was 5 mon hs (95% CI 3–8) s. 2 mon hs (95% CI 1.0—no eached) (p = 0.044) (Figu e 2A). The co esponding alues o OS we e 13 mon hs (95% CI 8.9–33) s. 3.2 mon hs (95% 2.0—no eached) (p = 0.0035), Figu e 1. 3-Hyd oxybu y a e and ace one as p ognos ic me aboli es in he Cox uni a ia e analysis (A,B) and mul i a ia e eg essions models (C–F) o p og ession- ee su i al (PFS) (A,C,E) and o e all su i al (OS) (B,D,F). Cance s 2025,17, 532 8 o 16 The op imal cu o s o se um 3OHB and ace one we e iden i ied wi h an ou come- o ien ed me hod (see Me hods) and we e se a 141 µ M and 40 µ M, espec i ely. Following he applica ion o he cu o , he PFS and OS Kaplan–Meie su i al cu es we e signi i- can ly di e en . Fo se um 3OHB, he median PFS was 5 mon hs (95% CI 3–8) s. 2 mon hs (95% CI 1.0—no eached) (p= 0.044) (Figu e 2A). The co esponding alues o OS we e 13 mon hs (95% CI 8.9–33) s. 3.2 mon hs (95% 2.0—no eached) (p= 0.0035), espec- i ely (Figu e 2B). Di e ences in PFS o se um ace one we e 5 mon hs (95% CI 3–9) s. 2 mon hs (95% CI 1.0—no eached) (p= 0.0054) (Figu e 2C) and 15 mon hs (95% CI 9–33) s. 2.5 mon hs (95% CI 1.8—no eached) (p= 0.00014) o OS, espec i ely (Figu e 2D). Signi ican di e ences we e also obse ed in uni a ia e Cox eg essions ollowing he applica ion o he cu o . Fo se um 3OHB, he HR o PFS was 2.21, while he HR o OS was 3.21. Fo se um ace one, he HR o PFS was 2.76 and he HR o OS was 3.71 (Supplemen a y Figu e S4). Cance s 2025, 17, x FOR PEER REVIEW 9 o 18 espec i ely (Figu e 2B). Diffe ences in PFS o se um ace one we e 5 mon hs (95% CI 3– 9) s. 2 mon hs (95% CI 1.0—no eached) (p = 0.0054) (Figu e 2C) and 15 mon hs (95% CI 9–33) s. 2.5 mon hs (95% CI 1.8—no eached) (p = 0.00014) o OS, espec i ely (Figu e 2D). Signi ican diffe ences we e also obse ed in uni a ia e Cox eg essions ollowing he applica ion o he cu off. Fo se um 3OHB, he HR o PFS was 2.21, while he HR o OS was 3.21. Fo se um ace one, he HR o PFS was 2.76 and he HR o OS was 3.71 (Supplemen a y Figu e S4). Figu e 2. Kaplan–Meie su i al cu es. (A,B), p og ession- ee su i al (PFS) and o e - all su i al (OS) o he cu off o 3-hyd oxybu y a e (3OHB); (C,D) PFS and OS o he cu off o ace one. 3.4. Nomog am o R/R DLBCL Risk S a i ica ion Nomog ams o isk s a i ica ion in pa ien s wi h R/R DLBCL based on se um me- aboli es 3OHB and ace one showed a high p edic i e pe o mance, sligh ly mo e a o a- ble o 3OHB, wi h an AUC o 0.856 o 6-mon h PFS (Figu e 3A,B) and 0.844 o 12-mon h OS (Figu e 3C,D). Se um ace one showed a simila p edic i e pe o mance, wi h an AUC o 0.840 o 6-mon h PFS (Figu e 3D,E) and 0.830 o 12-mon h OS (Figu e 3F,G). A nom- og am ha included bo h 3OHB and ace one was also explo ed; howe e , i pe o med wo se (da a no shown) since bo h me aboli es showed a s a is ically signi ican co ela- ion ( = 0.76, p < 0.001) (Supplemen a y Figu e S5). Le us conside he OS nomog am o 3OHB o wo pa ien s: Pa ien 1 and Pa ien 2. Al hough bo h pa ien s ha e an IPI sco e o 4–5 (3 poin s) and a e bo h e ac o y (18 poin s), Pa ien 1 has a 3OHB concen a ion o 800 µM (50 poin s), whe eas Pa ien 2 has a 3OHB concen a ion o 50 µM (3 poin s). Hence, o he 3OHB OS nomog am, Pa ien 1 sco ed 71 poin s (<10% su i al p obabili y a 12 mon hs), while Pa ien 2 sco ed 24 poin s (30% su i al p obabili y a 12 mon hs). Figu e 2. Kaplan–Meie su i al cu es. (A,B), p og ession- ee su i al (PFS) and o e all su i al (OS) o he cu o o 3-hyd oxybu y a e (3OHB); (C,D) PFS and OS o he cu o o ace one. 3.4. Nomog am o R/R DLBCL Risk S a i ica ion Nomog ams o isk s a i ica ion in pa ien s wi h R/R DLBCL based on se um me abo- li es 3OHB and ace one showed a high p edic i e pe o mance, sligh ly mo e a o able o 3OHB, wi h an AUC o 0.856 o 6-mon h PFS (Figu e 3A,B) and 0.844 o 12-mon h OS (Figu e 3C,D). Se um ace one showed a simila p edic i e pe o mance, wi h an AUC o 0.840 o 6-mon h PFS (Figu e 3D,E) and 0.830 o 12-mon h OS (Figu e 3F,G). A nomo- g am ha included bo h 3OHB and ace one was also explo ed; howe e , i pe o med wo se (da a no shown) since bo h me aboli es showed a s a is ically signi ican co ela ion ( = 0.76 , p< 0.001 ) (Supplemen a y Figu e S5). Le us conside he OS nomog am o 3OHB o wo pa ien s : Pa ien 1 and Pa ien 2. Al hough bo h pa ien s ha e an IPI sco e o 4–5 Cance s 2025,17, 532 9 o 16 ( 3 poin s ) and a e bo h e ac o y (18 poin s), Pa ien 1 has a 3OHB concen a ion o 800 µM (50 poin s), whe eas Pa ien 2 has a 3OHB concen a ion o 50 µ M (3 poin s). Hence, o he 3OHB OS nomog am, Pa ien 1 sco ed 71 poin s (<10% su i al p obabili y a 12 mon hs), while Pa ien 2 sco ed 24 poin s (30% su i al p obabili y a 12 mon hs). Cance s 2025, 17, x FOR PEER REVIEW 10 o 18 Figu e 3. Nomog ams o R/R DLBCL isk s a i ica ion based on single me aboli es and hei espec i e ime-dependen ROCs. (A,B) 3-Hyd oxybu y a e (3OHB) nomog am o p og ession- ee su i al (PFS) and i s ROC; (C,D) 3OHB nomog am o o e all su i al (OS) and i s ROC; (E,F) ace one nomog am o PFS and i s ROC; (G,H) ace one nomo- g am o OS and i s ROC. 4. Discussion The p esen da a con ibu e o de ining he me abolomic p o ile linked o R/R- DLBCL pa ien s’ esponsi eness o ea men and su i al. To he bes o ou knowledge, no p e ious me abolomics s udy has been epo ed in his se ing. T ea men esponden s Figu e 3. Nomog ams o R/R DLBCL isk s a i ica ion based on single me aboli es and hei espec i e ime-dependen ROCs. (A,B) 3-Hyd oxybu y a e (3OHB) nomog am o p og ession- ee su i al (PFS) and i s ROC; (C,D) 3OHB nomog am o o e all su i al (OS) and i s ROC; (E,F) ace one nomog am o PFS and i s ROC; (G,H) ace one nomog am o OS and i s ROC. Cance s 2025,17, 532 16 o 16 28. Feng, S.; Wang, H.; Liu, J.; AA, J.; Zhou, F.; Wang, G. Mul i-Dimensional Roles o Ke one Bodies in Cance Biology: Oppo uni ies o Cance The apy. Pha macol. Res. 2019,150, 104500. [C ossRe ] 29. Huang, D.; Li, T.; Wang, L.; Zhang, L.; Yan, R.; Li, K.; Xing, S.; Wu, G.; Hu, L.; Jia, W.; e al. Hepa ocellula Ca cinoma Redi ec s o Ke olysis o P og ession unde Nu i ion Dep i a ion S ess. Cell Res. 2016,26, 1112–1130. [C ossRe ] 30. Zhang, J.; Chen, B.; Zhang, C.; Zhu, M.; Fan, Z.; Li, L.; Wang, J.; Jin, J. The P ognos ic Value o Hyd oxybu y a e Dehyd ogenase in Di use La ge B Cell Lymphoma. iScience 2024,27, 110905. [C ossRe ] [PubMed] 31. Ca o, P.; Kishan, A.U.; No be g, E.; S anley, I.; Chapuy, B.; Fica o, S.B.; Polak, K.; Tonde a, D.; Gouna ides, J.; Yin, H.; e al. Me abolic signa u es unco e dis inc a ge s in molecula subse s o di use la ge B cell lymphoma. Cance Cell 2012,22, 547–560. [C ossRe ] [PubMed] 32. Penning, T.M.; Jonnalagadda, S.; T ippie , P.C.; Rižne , T.L.; Go esman, M. Aldo-Ke o Reduc ases and Cance D ug Resis ance. Pha macol. Re . 2021,73, 1150–1171. [C ossRe ] [PubMed] 33. Dedko a, E.N.; Bla e , L.A. Role o β -Hyd oxybu y a e, I s Polyme Poly- β -Hyd oxybu y a e and Ino ganic Polyphospha e in Mammalian Heal h and Disease. F on . Physiol. 2014,5, 101953. [C ossRe ] 34. Huang, C.K.; Chang, P.H.; Kuo, W.H.; Chen, C.L.; Jeng, Y.M.; Chang, K.J.; Shew, J.Y.; Hu, C.M.; Lee, W.H. Adipocy es P omo e Malignan G ow h o B eas Tumou s wi h Monoca boxyla e T anspo e 2 Exp ession ia β -Hyd oxybu y a e. Na . Commun. 2017,8, 14706. [C ossRe ] [PubMed] 35. Sko upa, A.; Po´nski, M.; Ciszek, M.; Cicho´n, B.; Klimek, M.; Wi ek, A.; Pakuło, S.; Boguszewicz, Ł.; Sokół, M. G ading o Endome ial Cance Using 1H HR-MAS NMR-Based Me abolomics. Sci. Rep. 2021,11, 18160. [C ossRe ] [PubMed] 36. Ba ano ico a, E.; Racay, P.; Zubo , P.; Smola , M.; Kudelo a, E.; Halaso a, E.; D o ska, D.; Danko a, Z. Ci cula ing Me aboli es in he Ea ly S age o B eas Cance We e No Rela ed o Cance S age o Sub ypes bu Associa ed wi h Ki67 Le el. P omising S a is ical Disc imina ion om Con ols. Mol. Cell P obes 2022,66, 101862. [C ossRe ] [PubMed] 37. Goui and, V.; Gicquel, T.; Lien, E.C.; Jaune-Pons, E.; Da Cos a, Q.; Fine i, P.; Me ay, E.; Duluc, C.; Maye s, J.R.; Audebe , S.; e al. Ke ogenic HMG-CoA Lyase and I s P oduc B-hyd oxybu y a e P omo e Panc ea ic Cance P og ession. EMBO J. 2022, 41, e110466. [C ossRe ] 38. Ni, Y.; Xie, G.; Jia, W. Me abonomics o Human Colo ec al Cance : New App oaches o Ea ly Diagnosis and Bioma ke Disco e y. J. P o eome Res. 2014,13, 3857–3870. [C ossRe ] [PubMed] 39. Zhang, L.; Jin, H.; Guo, X.; Yang, Z.; Zhao, L.; Tang, S.; Mo, P.; Wu, K.; Nie, Y.; Pan, Y.; e al. Dis inguishing Panc ea ic Cance om Ch onic Panc ea i is and Heal hy Indi iduals by 1H Nuclea Magne ic Resonance-Based Me abonomic P o iles. Clin. Biochem. 2012,45, 1064–1069. [C ossRe ] 40. Yang, D.O.; Xu, J.; Huang, H.; Chen, Z. Me abolomic P o iling o Se um om Human Panc ea ic Cance Pa ien s Using 1H NMR Spec oscopy and P incipal Componen Analysis. Appl. Biochem. Bio echnol. 2011,165, 148–154. [C ossRe ] [PubMed] 41. Al ai i, A.; Bahashwan, S.; Alsaadi, M.; Malhan, H.; Aqeel, A.; Al-Kahi y, W.; Almehda , H.; Qad i, I. Me abolic Bioma ke s in B-Cell Lymphomas o Ea ly Diagnosis and P edic ion, as Well as Thei In luence on P ognosis and T ea men . Diagnos ics 2022, 12, 394. [C ossRe ] [PubMed] 42. Al ai i, A.; Re ai, M.Y.; Alsaadi, M.; Bahashwan, S.; Malhan, H.; Al-Kahi y, W.; Dammag, E.; Ageel, A.; Mahza y, A.; Albihey i, R.; e al. Me abolomics: A New E a in he Diagnosis o P ognosis o B-Cell Non-Hodgkin’s Lymphoma. Diagnos ics 2023,13, 861. [C ossRe ] 43. Guai a-es e uelas, S.; Saa ed a-Ga cíaga cía, P.; Bosque , A.; Bo , J.; Gi ona, J.; Amiliano, K.; Rod íguez od íguez-Balada, M.; He as, M.; Masana, L.; Gum, J. Adipose-De i ed Fa y Acid-Binding P o eins Plasma Concen a ions A e Inc eased in B eas Cance Pa ien s. Oncologis 2017,22, 1309–1315. [C ossRe ] [PubMed] 44. Guai a-Es e uelas, S.; Gumà, J.; Masana, L.; Bo às, J. The Pe i umou al Adipose Tissue Mic oen i onmen and Cance . The Roles o Fa y Acid Binding P o ein 4 and Fa y Acid Binding P o ein 5. Mol. Cell Endoc inol. 2018,462, 107–118. [C ossRe ] 45. Guai a-Es e uelas, S.; Bosque , A.; Saa ed a, P.; Gumà, J.; Gi ona, J.; Lam, E.W.F.; Amillano, K.; Bo às, J.; Masana, L. Exogenous FABP4 Inc eases B eas Cance Cell P oli e a ion and Ac i a es he Exp ession o Fa y Acid T anspo P o eins. Mol. Ca cinog. 2017,56, 208–217. [C ossRe ] [PubMed] 46. Bonuccelli, G.; Tsi igos, A.; Whi ake -Menezes, D.; Pa lides, S.; Pes ell, R.G.; Chia a ina, B.; F ank, P.G.; Flomenbe g, N.; Howell, A.; Ma inez-Ou schoo n, U.E.; e al. Ke ones and Lac a e “Fuel” Tumo G ow h and Me as asis. Cell Cycle 2010,9, 3506–3514. [C ossRe ] Disclaime /Publishe ’s No e: The s a emen s, opinions and da a con ained in all publica ions a e solely hose o he indi idual au ho (s) and con ibu o (s) and no o MDPI and/o he edi o (s). MDPI and/o he edi o (s) disclaim esponsibili y o any inju y o people o p ope y esul ing om any ideas, me hods, ins uc ions o p oduc s e e ed o in he con en .