Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Translational Medicine: Patterns of Response to Antidepressant Treatment and their Implications"


Published on

This is a talk I gave at the IEEE Schenectady Section - 17 MAY Membership Meeting.

The mission of my depression research is to help people figure out what they need to help them get out of a depressed state. That is, finding out what is best for them, not what is best for their doctor, friends, therapist, or anyone else. Depression is now a global problem. In the past 15 years it has gotten worse. Depression is complex; it has a wide range of varying symptoms and degrees of intensity. It can be challenging to determine the best course of action, whether medical treatment is necessary, or which of the many treatments (drug and non-drug) is the best match. Many people who are depressed do not get the help they need, and many people receive medications when they are not necessary. My work aims to bring together tools, technology, scientific and medical data and patient experience to help address depression, both personally and globally.

Published in: Health & Medicine
  • Be the first to comment

Translational Medicine: Patterns of Response to Antidepressant Treatment and their Implications"

  1. 1. 1Translational Medicine: 
Patterns in Response to AntidepressantTreatment and their Implications"Joanne S. Luciano, PhD!The Tetherless World Constellation!Rensselaer Polytechnic Institute, Troy, NY!May 17, 2013!!!IEEE Schenectady Section!Sponsored by!The Engineering in Medicine & Biology Society (EBMS)!
  2. 2. 2Predictive Medicine, Inc. © 20102!Questions!Some people on antidepressants commitsuicide. Is it possible that theantidepressant drug can cause this tohappen? !!How can differential equations help us tounderstand what is going on?!
  3. 3. 33!Take Home Messages !!Depression is complex and personal.It requires unbiased multidisciplinaryapproach focused on the person.!!Depression is a global concern. TheWeb has global reach and needs tobe explored as a resource for helpingpeople with depression. !!Effectiveness may be the same, but!!patterns differ and order of symptomimprovement matters. !!
  4. 4. 4Predictive Medicine, Inc. © 20104!The$Economic$Burden$Source: The Healthy Thinking Initiative! is thehighest of the healthcare cost for business
  5. 5. 5Predic and 2006: Estimates for the U.S.Civilian Noninstitutionalized PopulationThe Five Most Costly Conditions
  6. 6. 66!The$Global$Burden$of$Depression$WHOs projections state that by2020 depression with be the 2ndlargest health burden worldwide,by 2030, it will be the largest.
  7. 7. 77!What is depression?Depression is a common mental disorder that presents withdepressed mood, loss of interest or pleasure, decreasedenergy, feelings of guilt or low self-worth, disturbed sleep orappetite, and poor concentration. Moreover, depression oftencomes with symptoms of anxiety. These problems can becomechronic or recurrent and lead to substantial impairments in anindividual’s ability to take care of his or her everydayresponsibilities.At its worst, depression can lead to suicide.Almost 1 million lives are lost yearly due to suicide, whichtranslates to 3000 suicide deaths every day. For every personwho completes a suicide, 20 or more may attempt to end his orher life (WHO, 2012)
  8. 8. 8Predictive Medicine, Inc. © 20108!Overview!•  Why we did this work - to improve quality of life for millionsof people suffering from depression!•  How we did it - used differential equations (“neuralnetwork”) to model and compare response to differentantidepressant treatments!•  What we found - different response patterns for the twotreatments - the order and timing of improvement ofsymptoms were different!•  What we think it means - improvement in selection oftreatment thereby reducing unnecessary costsand suffering. Potentially saving lives"
  9. 9. 9Predictive Medicine, Inc. © 20109!Overview!•  Why we did this work - to improve quality oflife for millions of people suffering fromdepression"•  How we did it - used differential equations (“neuralnetwork”) to model and compare response todifferent antidepressant treatments!•  What we found - different response patterns forthe two treatments - the order and timing ofimprovement of symptoms were different!•  What we think it means - improvement in selectionof treatment thereby reducing unnecessary costsand suffering. Potentially saving lives!
  10. 10. 10Predictive Medicine, Inc. © 201010!Translational Medicine!•  Rapid transformation of laboratory findings intoclinically focused applications !•  ‘From bench to bedside and back’!•  “and back” includes patients!!
  11. 11. 11Predictive Medicine, Inc. © 201011!HUGE PROBLEM!Characterized by persistent and pathological sadness,dejection, and melancholy!Prevalence (US)!!6% year (18 million)!!16% experience it in their lifetime!Cost !!44 Billion (1990)!Impact!!1% Improvement means (180, 000 people helped)!!1% Improvement means (440 million in savings)!
  12. 12. 12Predictive Medicine, Inc. © 2010Widespread!
  13. 13. 13Predictive Medicine, Inc. © 2010Treatment Choice Vague
No easy answer!
  14. 14. 14Predictive Medicine, Inc. © 201014!Overview!•  Why we did this work - to improve quality of life for millionsof people suffering from depression!•  How we did it - used differential equations(“neural network”) to model and compareresponse to different antidepressanttreatments!•  What we found - different response patterns for the twotreatments - the order and timing of improvement ofsymptoms were different!•  What we think it means - improvement in selection oftreatment thereby reducing unnecessary costs andsuffering. Potentially saving lives!
  15. 15. 1515!$$Research Goals!Illuminate recoverycourse(personalized)Properly diagnose and properlymatch patient with the best individualizedtreatment option available, includingnon-drug treatments
  16. 16. 16Predictive Medicine, Inc. © 201016!Today’s$talk$focuses$on:$ResponsetotreatmentTreatment Response Study!
  17. 17. 17Predictive Medicine, Inc. © 201017!Depression Background!•  Clinical Depression!•  Treatment!•  Symptom Measurement!•  No specific diagnosis!•  No specific treatment!
  18. 18. 18Predictive Medicine, Inc. © 201018!Clinical Data!Symptoms!! -HDRS (0-4 scale)!!Treatment!-Desipramine (DMI)!-Cognitive Behavioral Therapy (CBT)!!Outcome!! - Responders!
  19. 19. 19Predictive Medicine, Inc. © 201019!Hamilton 
Psychiatric Scale for Depression!Clinical Instrument standard measure in clinical trials. !Example of first three items of 21 items that measure individual!Symptom intensity.
  20. 20. 20Predictive Medicine, Inc. © 201020!Why Model?!!Easier to understand!Easier to manipulate!Easier to analyze!Recasting$the$problem$into$mathematical$terms$makes$it:$
  21. 21. 21Predictive Medicine, Inc. © 201021!Understanding Recovery
  22. 22. 22Predictive Medicine, Inc. © 201022!Understanding Recovery!
  23. 23. 23Predictive Medicine, Inc. © 201023!Depression Data!7 Symptom Factors " ""!Physical:! !E Sleep ! !! ! !M, L Sleep ! ! ! !! ! !Energy ! ! ! ! !!Performance: !Work & Interests ! ! ! !!Psychological: !Mood ! ! ! ! !! ! !Cognitions ! ! ! !! ! !Anxiety ! ! !!2 Treatments " !Cognitive Behavioural Therapy (CBT)! !! ! !Desipramine (DMI)!!"Clinical Data ! !Responders = improvement >= 50% on HDRS total! ! ! !N = 6 patient each study!! !6 weeks ! = 252 data points (converted to daily) !! ! ! each study (CBT and DMI)!!
  24. 24. 24Predictive Medicine, Inc. © 201024!Overview 
Recovery Model and Parameters!MEWMSESAC
  25. 25. 25Predictive Medicine, Inc. © 201025!Modelling 
Time to Response !
  26. 26. 26Predictive Medicine, Inc. © 201026!Modelling 
Treatment Effects!
  27. 27. 27Predictive Medicine, Inc. © 201027!Recovery Equation
(Luciano Model) 
  28. 28. 28Predictive Medicine, Inc. © 201028!Training the model!
  29. 29. 29Predictive Medicine, Inc. © 201029!Individual Patient Recovery Pattern and Error!Example Patient (CBT)!Fit of Model on for individual patient captures trends butnot entire pattern. Not good enough.!
  30. 30. 30Predictive Medicine, Inc. © 201030!Patient Group (CBT)!Recovery Pattern and Error!Model on data for patient treatment group capturesentire pattern. Good fit of Model to data.!
  31. 31. 31Predictive Medicine, Inc. © 201031!Overview!•  Why we did this work - to improve quality of life for millionsof people suffering from depression!•  How we did it - used differential equations (“neural network”)to model and compare response to different antidepressanttreatments!•  What we found - different response patternsfor the two treatments - the order and timingof improvement of symptoms were different"•  What we think it means - improvement in selection oftreatment - less trial and error !
  32. 32. 32Predictive Medicine, Inc. © 201032!Results
Optimized parameters specify model
Initial conditions predict pattern trajectory !MCWAEESMLS
  33. 33. 33Predictive Medicine, Inc. © 201033!Latency!
  34. 34. 34Predictive Medicine, Inc. © 201034!Mean ½ Reduction Time!CBT varies 3.7 wksDMI varies 1.8 wks
  35. 35. 35Predictive Medicine, Inc. © 201035!Direct Effect of Treatment!
  36. 36. 36Predictive Medicine, Inc. © 201036!Direct Treatment Intervention Effect!
  37. 37. 37Predictive Medicine, Inc. © 201037!Treatment Effects 
and Interaction Effects!CBTSequentialDMI(delayed)CONCURRENTDMI:•  Interactions > 2x•  Loops
  38. 38. 38Predictive Medicine, Inc. © 2010Order and Time asymptom improvesare both different 
!Different Response Patterns !for Different Treatment!CBTDMICBT (talk: no drugs) 
DMI (drug: tricyclic antidepressant)!This is important because it showshow an antidepressant medicationcould lead to a suicide.

By giving a suicidal patient DMI, youcould increase the patients energybefore the suicidal thoughts improve.This could give them the energy toact on those suicidal thoughts.!
  39. 39. 39Predictive Medicine, Inc. © 201039!Overview!•  Why we did this work - to improve quality of life for millionsof people suffering from depression!•  How we did it - used differential equations (“neural network”)to model and compare response to different antidepressanttreatments!•  What we found - different response patterns for the twotreatments - the order and timing of improvement ofsymptoms were different!•  What we think it means - improvementin selection of treatment therebyreducing unnecessary costs andsuffering. Potentially saving lives.!
  40. 40. 40Predictive Medicine, Inc. © 201040!Conclusions!•  A neural network model is capable of predictingand describing recovery patterns in depression!•  We can do better than trial and error treatmentprotocols, which are still the norm today!!•  Recovery patterns differ by treatment!•  Cognitive Behavioural Therapy!is sequential!•  Desipramine!is concurrent (after delay)!•  Recovery patterns provide insights to patientresponse that can inform treatment choices!
  41. 41. 41Predictive Medicine, Inc. © 201041!Limitations!Model:!•  Assumes symptoms interact!•  Assumes treatment acts directly!•  Permanent vs. transient!•  Causal vs. sequential!•  Statistical fluctuations not handled!Study:!•  CBT measurement intervals vary!•  Small sample size!•  Initial 6 weeks of CBT (entire=16)!•  Finer resolution of measurements (2-3/day)!
  42. 42. 42Predictive Medicine, Inc. © 201042!Thank$you!$
  43. 43. 43Predictive Medicine, Inc. © 201043!Backup$Slides$
  44. 44. 44Predictive Medicine, Inc. © 2010Recovery Model
  45. 45. 45Predictive Medicine, Inc. © 201045!
  46. 46. 46Predictive Medicine, Inc. © 201046!
  47. 47. 47Web Observatories
@ Rensselaer WSRC!At RPI WSRC, our observatoriesbuild and monitor new tools andprocesses that are needed toaddress the Web s complexityand multifaceted nature.
  48. 48. 48Automating Health Data"Health Web Science is an emerging subfield of WebScience which looks to understand the increasingamounts of data from the health and life sciencefields.In order to advance Health Web Science, oneapproach is in automating health data.RPI TWC builds tools to automate datainteroperability across the web(and therefore the world).
  49. 49. 49Department of Health and Human Services!                 Developer Challenge"6In June 2012, HHS issued the first of its seven challenges calling fordevelopers to make high value health data more accessible toentrepreneurs, researchers, and policy makers in the hopes of betterhealth outcomes for all. !!A group from RPI TWC won first place in the competition, by usingsemantic technologies and in-house developed software, such ascsv2rdf4lod, LODSPeaKr, Farrah and DataFAQS.!HHS wanted Metadata!!"... application of existing voluntary consensusstandards for metadata common to all opengovernment data"!"RPI TWC submitted:!• DCAT - W3C Data Catalog! Version controlled on github.! Extracted from their CKAN as input toconverter.!• VoID - W3C Vocabulary of Interlinked Data! Organized datasets by source, dataset,version.! Provided links to data dumps, Linksets to LOD.!• PROV - W3C Provenance InterchangeModel! Captured during CKAN extraction, retrieval,conversion, and publishing.!• Dublin Core Metadata Terms! Annotated subjects based on descriptions.!HHS wanted Classification""...classify datasets in our growing catalog,creating entities, attributes and relations thatform the foundations for better discovery,integration...""!RPI TWC presented:"• Bottom-up vocabulary and entity reuse◦ Vocabulary created for each dataset◦ Enhanced datasets shifted to reusevocabulary and entities from other datasets.◦ Three stub vocabularies for top-level reuse.• NCBO (Nat. Center for Biomedical Ont.)Annotations◦ annotator/ SADI service◦ data/source/bioontology-org/annotator-description-subject/version/retrieve.shHHS wanted Liquidity"!"new designs ... that form the foundations for ... liquidity"!!RPI TWC provided: 2B triples among 1M URIs!• Dataset Linked Data! Machine and Human views (via conneg)! Faceted search of datasets!• Dataset dumps (.ttl.gz)! For each dataset, and for the whole thing.!Dataset query (!
  50. 50. 50 6Foundations and Trends
Series!Foundations and Trends in!Health Web Science

NOW Publishers
(available fall 2013)!!!!!prepubication: to order, send emailto
  51. 51. 51Predictive Medicine, Inc. © 201051!Acknowledgements$Sam$Levin$Dan$Levine$Dan$Bullock$Ennio$Mingola$Michiro$Negishi$Jacqueline$Sampson$Larry$Hunter$Rick$Lathrop$Larrie$Hutton$Tim$Clark$$Jeremy Zucker!Alan Ruttenberg!Jonathan Rees!Robert Stevens!Phil Lord!Alan Rector!Andy Brass!Paul Fisher!Carole Goble!George Church!Matt Temple!ChristopherBrewsterEric Neumann!Chris Sander!Mike Cary!!ME Patti!Mark Musen!Zak Kohane!Brian Athey!David States!!!RPI TWC especially:!Peter Fox!Jim Hendler!Deborah McGuinness!Yuezhang Xiao!Brendan Ashby!Zach Jablons!Aishwarya Venkatakrishnan !!!!
  52. 52. 52Joanne S. Luciano, BS, MS, PhD!Academic:!!!!Rensselaer Polytechnic Institute, Troy, NY!!Consulting:!!Predictive Medicine, Inc., Belmont, MA!Predictive Medicine, Inc. © 2010CV Background slides...!
  53. 53. 53Predictive Medicine, Inc. © 201053!Spanning$disciplines$Emerging$disciplines$Diagnosis & TreatmentDepressionTranslationalMedicineHuntington sDiseaseMedicine LifeSciencesInformation SystemsNeuroscienceMolecularBiologyClinical PracticeSigns and SymptomsBiochemistryMathematicalModelsComputersimulationOntologyGeneticsGenomicsAnatomyMachine LearningSemantic WebInfluenzaResearchFindingsBioinformaticsElectronicMedicalRecordsClinicalResearchDiabetes
  54. 54. 54Timeline!(earlier work: 10 years in Software Research & Development and Product Development)!20091993World Congress onNeural Networks,July 11-15, 1993,Portland, Oregon SIGMental Function andDysfunctionSam LevinJackie Samson,Mc Lean HospitalDepressionResearch1996199520081994Patents Soldto AdvancedBiologicalLaboratoriesBelgiumPatents Offered atOcean TomoAuction Chicago, ILUS Patent No.6,317,73AwardedUS PatentsNo. 6,063,028Awarded20012000PhDThesis ProposalApprovedWorkshop Neural Modeling ofCognitive and Brain DisordersBioPAX?Linked DataW3C HCLSBioDASHEPOS2006EMPWRPoster PresentedISMB 1997PSB 199819972010Rensselaer(RPI)2011 20122013U PittGreg SiegleCollaborationYuezhangXiaoMaster’sThesis(RPI)Brendan AshbyMaster’sThesis (RPI)Center for"Multi-disciplinaryResearchand"Depression"Treatment"Selection"""
  55. 55. 55Predictive Medicine, Inc. © 2010Neural$Modeling$of$Depression$1996 Luciano, J., Cohen, M. Samson, J.”Neural Network Modeling of UnipolarDepression,” Neural Modeling ofCognitive and Brain Disorders, WorldScientific Publishing Company, eds. J.Reggia and E. Ruppin and R. Berndt.Book cover; chapter pp 469-483.Luciano Model highlighted on book coverWorkshop 1995Book 1996
  56. 56. 56Predictive Medicine, Inc. © 201056!BioPathways$Consortium$$$$BioPAX$$$$$$$W3C$Semantic$Web$for$Health$Care$and$Life$Sciences$(HCLSIG)$Establishing$$Communities$of$Interest/Practice$
  57. 57. 57BioPAX$S$Enabling$$Cellular$Network$Process$Modeling$MetabolicPathwaysMolecularInteractionNetworksSignalingPathwaysGene RegulatoryNetworksGlycolysis Protein-Protein Apoptosis TFs in E. coli
  58. 58. 58Predictive Medicine, Inc. © 201058!27 October 2008EMPWR$$$Collaboration$with$Manchester,$UK$Use$instanceStore$to$reason$over$BioPAX$formatted$(OWL)$pathway$data$• Goal:$discover$new$scientific$facts$• Method:$Utilize$power$of$reasoners$and$OWL$through$coupling$BioPAX$data$and$Manchester$Technology$• Results:$BioPAX$semantics$lacking$thus$had$to$educate$BioPAX$community$and$courseScorrect$initiative$• Extending$BioPAX$to$enable$the$computational$exploration$
  59. 59. 59Predictive Medicine, Inc. © 201059!Diabetes:$Understanding$the$role$of$risk$factors$$in$insulin$resistance$Figure: Integration of genomic and proteomic/metabolomic data (text boxes shaded in gray) proposed for current project. We hypothesize that diabetes risk factors result in alteredgene and protein expression in skeletal muscle and adipose tissue (genomic data), leading to insulin resistance and inflammation. This, in turn, results in abnormal tissue function, asindicted by accumulation of long-chain fatty acyl CoA and oxidative damage (proteomic and metabolomic data), further insulin resistance and beta-cell failure, and ultimately to type 2diabetes
  60. 60. 60Predictive Medicine, Inc. © 201060!Enhance$capability$Cell Designer model of adipose tissue cell. Add geneexpression, standard metadata terms (BioPAX, GenBank)!Use with expression data constrained by proteomic data!towards target ID, biomarker ID, patient population ID!
  61. 61. 61Predictive Medicine, Inc. © 201061!27 October 2008Licensing$Opportunities$Available$US Patent No. 6,063,028 May 2000AUTOMATED TREATMENTSELECTION METHODUS Patent No. 6,317,731 Nov 2001METHOD FOR PREDICTING THEOUTCOME OF A TREATMENT2009$Sold$to$Advanced$Biomedical$Labs$(Luxembourg)$
  62. 62. 62Predictive Medicine, Inc. © 201062!Take$Home$Message $$We$are$shortening$the$time$$and$tightening$the$loop$between$research$and$practice,$however….$We$need$to$do$better$S$15$years$+$is$too$long,$way$too$long$for$depressed$people$to$be$suffering$needlessly.$We$must$engage$all$stakeholders$and$give$citizens$power$over$their$health$data$
  63. 63. 63Predictive Medicine, Inc. © 201063!Research$and$Practice$!   Computational$modelers$construct$in#silico$representations$of$organic$phenomena$$!   Basic$researchers$construct$in#vitro#!   Clinical$Researcher’s$conduct$in#vivo$studies$on$patient$populations$!   Clinical$practioners$apply$the$results$of$clinical$research$