qiprofile/app/javascripts/breast.coffee

Quantitative Imaging Profile ( QiPr) web application

define ['angular', 'lodash', 'helpers'], (ng, _) ->
  breast = ng.module 'qiprofile.breast', ['qiprofile.helpers']

  breast.factory 'Breast', ['ObjectHelper', (ObjectHelper) ->
    # The cancer stages correspond to TNM scores, assuming no
    # metastasis (M0).
    #
    # TODO - extend this table to account for the T and N suffixes per
    # the reference document mentioned below.
    STAGES = [
     ['1A', '2A', '3A', '3C']
     ['1A', '2A', '3A', '3C']
     ['2A', '2B', '3A', '3C']
     ['2B', '3A','3A', '3C']
     ['3B', '3B','3B', '3C']
    ]

    Grade:
      RANGES: [[3..5], [6..7], [8..9]]
      SCORES: ['tubularFormation', 'mitoticCount', 'nuclearPleomorphism']

    # @returns the sorted stage values
    stageExtent: ->
      _.chain(STAGES).flatten().union(['4']).uniq().sort().value()

    # Returns the cancer stage.
    #
    # 1f metastasis exists (M1), then the stage is 4.
    # Otherwise, the stage is determined by T and N scores as
    # defined in the tumor type factory STAGES associative
    # lookup table.
    #
    # @param tnm the TNM object
    # @returns the cancer stage object, as described in tnm.coffee
    #    stage
    stage: (tnm) ->
      # M1 => stage IV.
      # TODO - shouldn't this be '4'? Is there a test case for this?
      #   Since the value is romanized prior to display, this bug might
      #   be hidden. Same with the Sarcoma stage.
      if tnm.metastasis
        return 'IV'
      # The T and N scores.
      # TODO - factor in the size suffix.
      t = tnm.size.tumorSize
      n = tnm.lymphStatus

      # Lookup (T, N) in the stage table.
      # If not found, then throw an error.
      find = (table, value) ->
        result = table[value] or
          throw new ReferenceError("Unsupported #{ tnm.tumorType }" +
                                   " TNM: #{ ObjectHelper.prettyPrint(tnm) }")

      _.reduce([t, n], find, STAGES)
    

    # Returns the cancer recurrence score. This score is calculated from
    # a genetic expression assay according to algorithm in Figure 1 of
    # the following paper:
    #
    #   Paik, et al., 'A Multigene Assay to Predict Recurrence of
    #   Tamoxifen-Treated, Node-Negative Breast Cancer',
    #   N Engl J Med 2004; 351:2817-2826
    #   (http://www.nejm.org/doi/full/10.1056/NEJMoa041588)
    #
    # 1f metastasis exists (M1), then the stage is 4.
    # Otherwise, the stage is determined by T and N scores as
    # defined in the tumor type factory STAGES associative
    # lookup table.
    #
    # @param tnm the TNM object
    # @returns the cancer stage object, as described in tnm.coffee
    #    stage
    recurrenceScore: (assay) ->
      her2Unscaled = (0.9 * assay.her2.grb7) + (0.1 * assay.her2.her2)
      her2 = Math.max(8, her2Unscaled)
      erUnscaled = (0.8 * assay.estrogen.er) + (1.2 * assay.estrogen.pgr) +
                   assay.estrogen.bcl2 + assay.estrogen.scube2
      er = erUnscaled / 4
      proliferationUnscaled = (assay.proliferation.survivin + assay.proliferation.ki67 +
                              assay.proliferation.mybl2 + assay.proliferation.ccnb1 +
                              assay.proliferation.stk15) / 5
      proliferation = Math.max(6.5, proliferationUnscaled)
      invasion = (assay.invasion.ctsl2 + assay.invasion.mmp11) / 2

      # The unscaled score.
      recurrenceUnscaled = (0.47 * her2) - (0.34 * er) + (1.04 * proliferation) +
                           (0.10 * invasion) + (0.05 * assay.cd68) -
                           (0.08 * assay.gstm1) - (0.07 * assay.bag1)
      
      # Guard against missing values.
      if isNaN(recurrenceUnscaled)
        return null
      
      recurrenceScaled = Math.round(20 * (recurrenceUnscaled - 6.7))
      
      # Return the score fit to the range [0, 100].
      Math.max(0, Math.min(recurrenceScaled, 100))


    # Calculates the Residual Cancer Burden index and class as described in:
    #   JCO 25:28 4414-4422 <http://jco.ascopubs.org/content/25/28/4414.full>
    #
    # @param tumor the tumor object
    # @returns the RCB object extended with index and class properties
    residualCancerBurden: (tumor) ->
      # @param extent the tumor extent {length, width, depth} REST object
      # @param rcb the RCB REST object
      # @returns the RCB index
      rcbIndex = (extent, rcb) ->
        # The bidimensional tumor size metric.
        size = Math.sqrt(extent.length * extent.width)
        # The overall tumor cellularity.
        overall = rcb.tumorCellDensity / 100
        # The in situ cellularity.
        inSitu = rcb.dcisCellDensity / 100
        # The invasive carcinoma proportion.
        invasion = (1 - inSitu) * overall
        # The RCB index invasion component.
        invasionFactor = 1.4 * Math.pow(invasion * size, 0.17)
        # The RCB index positive node component.
        posNodeFactor = 1 - Math.pow(0.75, rcb.positiveNodeCount)
        # The base of the RCB index node component.
        nodeBase =  4 * posNodeFactor * rcb.largestNodalMetastasisLength
        # The RCB index node component.
        nodeFactor = Math.pow(nodeBase, 0.17)

        # The RCB index is the sum of the invasion and node components.
        invasionFactor + nodeFactor

      # @param index the calculated RCB index value
      # @returns the RCB class based on RCB index cut-offs
      rcbClass = (index) ->
        if index == 0
          return 0
        else if index < 1.36
          return 1
        else if index < 3.28
          return 2
        else
          return 3

      # If the RCB object exists, then return it extended with the
      # index and class properties.
      rcb = tumor.rcb
      if rcb?
        rcb.index = rcbIndex(tumor.extent, tumor.rcb)
        rcb.class = rcbClass(rcb.index)
        rcb
  ]