111111111111111111111111111111111111111111111111111111111111111111111111111
`USOO5320102A
`Patent Number:
`Date of Patent:
`
`[11]
`
`[45]
`
`5,320,102
`Jun. 14, 1994
`
`United States Patent
`Paul et aI.
`
`[19]
`
`[54]
`
`[75]
`
`METHOD FOR DIAGNOSING
`PROTEOGLYCAN DEFICIENCY IN
`CARTILAGE BASED ON MAGNETIC
`RESONANCE IMAGE (MRI)
`
`Inventors: Pradip K. Paul, Cranford; Mukundrai
`K. Jasani, Berkeley Heights;
`Elizabeth O'Byrne, Millburn;
`Douglas Wilson, Scotch Plains;
`Ashok Rakhit, Berkeley Heights, all
`of N.J.
`
`[73]
`
`Assignee: Ciba-Geigy Corporation, Ardsley,
`N.Y.
`
`[21]
`
`App1. No.: 978,511
`
`[22]
`
`Filed:
`
`Nov. 18, 1992
`
`[51]
`[52]
`[58]
`
`[56]
`
`Int. Cl.s
`U.S. Cl.
`Field of Search
`
`A61B 5/055
`128/653.2
`128/653.2; 436/173
`
`References Cited
`U.S. PATENT DOCUMENTS
`5,206.023
`4/1993 Hunziker
`
`424/423
`
`OTHER PUBLICAnONS
`"Magnetic Resonance Imaging Reflects Cartilage Pro(cid:173)
`teoglycan Degradation in the Rabbit Knee", Pradip K.
`Paul et al.. Skeletal Radio1., (1991), 20:31-36.
`"Intevertebral Disk: Normal Age-related Changes in
`
`MR Signal Intensity)", Lowel A. Sether etal., Neurora(cid:173)
`diology, Nov. 1990, pp. 385-388.
`"RSNA'89 Scientific Program", Radiological Society
`of North America, Nov. 26-Dec. 1, Chicago, p. 436.
`"MRI Evaluation of Early Degenerative Cartilage Dis(cid:173)
`ease by a Three-dimensional Gradiaent Echo Se(cid:173)
`quence", K. Gliickert et al., Tissue Characterization in
`MR Imaging, pp. 185-192.
`"Articular Cartilage: Correlation of Histologic Zone
`with Signal Intensity at MR Imaging", Jean M. Modi et
`aI., Radiology, 1991, pp. 853-855.
`"Structure, Function, and Degeneration of Bovine Hy(cid:173)
`aline Cartilage: Assessment with MR Imaging in Vi(cid:173)
`tro!", Radiology, 1989, pp. 495-499.
`"Anatomical Changes and Patheogenesis of OA in
`Man, with Particular Reference to the Hip and Knee
`Joints", D. L. Gardner et al., pp. 22-48.
`"Biochemical Changes in Human Osteoarthrotic Carti(cid:173)
`lage", M. T. Bayliss, pp. 50-56.
`Primary Examiner-Ruth S. Smith
`Attorney, Agent, or Firm-Wenderoth, Lind & Ponack
`[57]
`ABSTRACT
`Proteoglycan deficiency in articular cartilage is diag(cid:173)
`nosed based on quantified signal intensities of pixels of a
`magnetic resonance image (MRI) extending across a
`depth of the articular cartilage. A pattern of the thus
`quantified signal intensities is indicative of proteoglycan
`distribution across the cartilage depth.
`
`24 Claims, 18 Drawing Sheets
`
`RECEIVE MR IMAGE DATA
`
`!
`
`DISPLAYING MR IMAGE
`
`!
`
`MAGNIFY AREA OF
`INTEREST
`~
`RECORD MR SIGNAL
`INTENSITY OF INDIVIDUAL
`PIXELS OF CARTILAGE
`
`!
`
`SIGNAL INTENSITY
`AVERAGING AND
`INTERPOLATION
`
`!
`
`PLOT MR SIGNAL
`INTENSITY VS.
`CARTILAGE DEPTH
`
`-1-
`
`Smith & Nephew Ex. 1014
`IPR Petition - USP 7,534,263
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 1 of 18
`
`5,320;102
`
`FIG. 1
`
`104
`
`~--105
`
`·W...,-~~
`
`106
`
`107
`
`108
`109
`110
`1----111
`112
`
`113
`
`115
`
`-2-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 2 of 18
`
`5,320,102
`
`FIG. 2
`
`RECEIVE MR IMAGE DATA
`
`DISPLAYING MR IMAGE
`
`MAGNIFY AREA OF
`INTEREST
`
`RECORD MR SIGNAL
`INTENSITY OF INDIVIDUAL
`PIXELS OF CARTILAGE
`
`SIGNAL INTENSITY
`AVERAGING AND
`INTERPOLATION
`
`PLOT MR SIGNAL
`INTENSITY VS.
`CARTILAGE DEPTH
`
`-3-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 3 of 18
`
`5,320,102
`
`FIG. 3
`
`FIG. 4(0)
`
`-4-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 4 of 18
`
`5,320;102
`
`FIG. 4(b)
`
`FIG. 4(c)
`-"-_.'._~ --=
`
`-5-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 5 of 18
`
`5,320,102
`
`FIG.5(a)
`350 -::r----------..,
`300
`250
`INTEN- 200
`SITY 150
`
`100
`
`50 0 1 2 3 4 5 6 7 8
`SURFACE
`SUBJECT 1
`BONE
`......AVG LATTIB INT .... AVG MEDTIB INT
`
`FIG.5(d)
`350-::r--------,
`300
`250
`INTEN· 200
`SITY 150
`100
`50~-----_.....I
`7 8
`0123456
`BONE
`SURFACE
`SUBJECT 4
`
`FIG.5(b)
`350 - , - - - - - - - - - - . . ,
`300-
`250-
`INTEN-200_ ~
`SITY 150 _
`""'11
`
`100-
`
`'
`
`FIG.5(e)
`350.....----------..,
`300-
`250- ~
`INTEN- 200~
`SITY 150~
`100-
`
`50 0 1 2 3, 4 5 6
`SURFACE
`SUBJECT 2
`
`7 8
`BONE
`
`50 0 1 2 3 4 5 6
`SURFACE
`SUBJECT 5
`
`7 8
`BONE
`
`FIG.5(c)
`350-.-----------.
`300
`250
`INTEN- 200
`SITY 150
`
`FIG.5(f)
`350......----------..,
`300-
`250- ~
`INTEN- 200-
`.
`SITY 150-
`
`100
`
`100-
`
`50 0 1 2 3 4 5 6
`SURFACE
`SUBJECT 3
`
`7 8
`BONE
`
`50 0 1 2 3 4 5 6 7 8
`SURFACE
`BONE
`MEAN FOR ALL SUBJECTS
`
`-6-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 6 of 18
`
`5,320,102
`
`FIG.6(a)
`350-,----------..
`300
`250
`INTEN- 200
`SITY 150
`100
`50 - ' - - - - - - - - - - '
`o 1 2 3 4 5 678
`SURFACE
`SUBJECT 1
`BONE
`-+- AVG LAT FEM INT
`AVG MED FEM INT
`
`FIG.6(d)
`
`350~-------,
`300
`250
`INTEN-200
`SITY 150
`100
`
`50 0 1 2 3 4 5 6
`SURFACE
`SUBJECT 4
`
`7 8
`BONE
`
`FIG.6(b)
`350 - , - - - - - - - - - - ,
`300-
`250(cid:173)
`INTEN- 200(cid:173)
`SITY 150..;
`
`FIG.6(e)
`350 - r - - - - - - - - - ,
`300
`250
`INTEN- 200
`SITY 150
`
`100-
`
`100
`
`50 0 1 2 3 4 5 6
`SURFACE
`SUBJECT 2
`
`7 8
`BONE
`
`50 0 1 2 3 4 5 6
`SURFACE
`SUBJECT 5
`
`7 8
`BONE
`
`FIG.6(c)
`350 - , - - - - - - - - - - ,
`300- ~
`250-
`,
`..
`INTEN- 200(cid:173)
`SITY 150-
`100-
`
`FIG.6(f)
`350 - : : r - - - - - - - - ,
`300
`250
`INTEN- 200
`SITY 150
`100
`
`50 0 1 2 3 4 5 6
`SURFACE
`SUBJECT 3
`
`7 8
`BONE
`
`50 0 1 2 3 4 5 6 7 8
`SURFACE
`BONE
`MEAN FOR ALL SUBJECTS
`
`-7-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 7 of 18
`
`5,320;102
`
`WU
`
`<C
`U-
`e:
`::>
`en
`
`0
`
`.....
`
`C")
`
`en
`--J
`
`WXa
`
`:
`W
`-e- CJ
`::s
`l-e:
`<C
`U
`
`L(')
`
`coW
`
`Z~
`
`0
`.....
`L(')
`
`L(')
`.....
`N
`
`0
`.....
`0
`
`L(')
`
`"
`
`o
`L(')
`
`L(')
`N
`
`L(')
`
`"
`.....
`>-
`!::::
`enz
`-
`Z
`
`WI
`
`L(')
`
`0
`N
`0
`0
`.....
`~
`0
`0
`.....
`0
`0
`"
`co
`+ + +
`
`r-,
`•
`i:
`
`~L
`
`0
`0
`L(')
`L(')
`NO " L(')
`N
`N
`C")
`C")
`
`L(')
`N
`N
`
`0
`0
`N
`
`-8-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 8 of 18
`
`5,320,102
`
`w
`za
`colD
`
`LO
`
`Cf')
`...J
`
`WX
`
`0:
`w
`~C!)«
`l-e:«
`
`...J
`
`<..)
`
`w
`w
`e:
`C!)
`w
`0
`a
`Ct)
`
`w
`W
`e:
`C!)
`w
`0
`a
`LO
`
`ui
`w
`e:
`C!)
`w
`0
`.....
`LO
`+ + +
`
`co
`•
`C)
`Li:
`
`Ct)
`
`w<
`
`..)«au..e:
`:::::>a:
`
`aL
`
`O
`
`a
`(j)
`
`a
`co
`
`a
`l'--
`
`a
`a
`.....
`>-I-
`
`a
`.....
`
`Ct)
`
`a
`.....
`C\.l
`
`a
`.....
`.....
`
`aL
`
`O.....
`
`Cf')
`
`ZWI
`
`-
`Z
`
`-9-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 9 of 18
`
`5,320;102
`
`I-+- COLlAGEN I
`
`I
`I-+- WATER
`~~
`
`I-+- PROTEOGLYCAN I
`
`I
`7
`BONE
`
`I
`6
`
`I5
`
`I4
`
`I3
`
`I
`1
`SURFACE
`
`I
`2
`
`80
`75
`
`70
`
`% DRY
`WEIGHT 65
`60
`55
`50
`75
`
`%WET
`WEIGHT 70
`
`65
`80
`
`75
`
`70
`
`%
`EXTRACTED 65
`
`60
`
`55
`
`FIG. 9
`
`-10-
`
`

`

`""'"
`..,..
`
`0N
`
`W
`N
`0
`
`..,.
`01
`
`00
`
`~
`
`...
`
`0
`0.
`
`~
`
`ll
`00
`
`~t
`
`tll....
`
`,fA
`\C
`
`C
`~\
`
`~,fA
`
`~
`
`INTENSITY
`
`SIGNAL
`
`~=:
`
`=
`
`tll
`
`=
`~
`~~n
`
`>
`
`IotJ
`0 •
`
`~•0
`
`40
`
`I
`
`30
`
`I
`
`L1000
`
`1100
`
`1200
`
`I1300
`
`1400
`
`1500
`
`600
`
`1700
`
`r1
`
`I1800
`
`SLICE-MICRONSX100
`
`20
`
`I
`
`10
`
`I
`
`I
`
`d-
`
`0-1
`
`II
`
`1-1
`
`\
`
`J(
`
`~
`
`RATIO
`OHPRO2
`GAG-S04/
`
`1900
`
`2000
`
`,2100
`
`•GAG-S04/0HPRORATIOL
`
`oSIGNALINTENSITY
`
`r\\
`
`/
`
`FIG.10
`
`3J
`
`4-,
`
`-11-
`
`

`

`I-to
`
`0N
`
`'It
`
`U1
`
`'It
`
`W
`N0
`
`QC)
`
`0"
`
`....
`""
`.... ....
`....
`
`1:1'"
`rJ:j
`
`~~
`
`~~
`
`==~
`
`C-4
`
`.... \C
`....
`
`\C
`
`~
`
`=f
`=
`
`f""to-
`n>
`""to-
`"'C
`rJ1 •
`c:: •
`
`0.0242
`0.0045
`-71.11
`13.54
`33.17
`160.91
`
`146.50
`182.20
`120.65
`172.85
`209.35
`133.90
`
`M-WEQUIV0.0001
`T-TEST:0.001
`%DIFF:
`-80.73
`STDERROR:14.38
`STD:35.23
`AVERAGES:152.33
`
`145.50
`115.30
`129.40
`187.80
`203.45
`132.55
`
`1
`5
`3
`4
`2
`6
`
`TIBIAL
`
`RA&NORMALCOMPARISONS{
`
`27.29
`72.21
`46.49
`22.45
`26.05
`18.80
`17.75
`10.95
`209.90
`19.50
`
`10.80
`28.58
`29.35
`19.35
`19.60
`17.55
`21.80
`15.30
`94.00
`17.85
`
`STDERROR:
`STD:
`AVERAGES:
`
`1
`4
`2
`6
`5
`3
`7
`
`SUBJECTMEDIALLATERAL
`
`NORMAL
`
`SUBJECTMEDIALLATERAL
`
`RA
`
`0.0005
`0.0001
`-77.32
`11.50
`28.18
`179.73
`
`192.54
`186.70
`134.80
`196.35
`210.70
`157.30
`
`M-WEQUIV:0.0076
`T-TEST:0.0012
`%DIFF:
`-73.01
`12.88
`STDERROR:
`31.54
`STD:
`159.82
`AVERAGES:
`
`193.95
`159.07
`135.10
`197.85
`118.30
`154.65
`
`1
`5
`3
`4
`2
`6
`
`FEMORAL
`
`RA&NORMALCOMPARISONS{
`
`FIG.11-:
`
`22.18
`17.91
`58.69
`47.39
`43.13
`40.76
`21.60
`25.15
`24.80
`30.60
`22.75
`22.90
`21.15
`21.50
`17.80
`14.65
`176.10
`147.70
`17.70.22.80
`
`STDERROR:
`STD:
`AVERAGES:
`
`1
`4
`2
`6
`5
`3
`7
`
`COMPARISONOFPEAKINTENSITYINFEMORALANDTIBIALCARTILAGESOF7RA
`
`PATIENTSWITH5NORMALVOLUNTEERS
`
`SUBJECTMEDIALLATERAL
`
`NORMAL
`
`SUBJECTMEDIALLATERAL
`
`RA
`
`-12-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 12 of 18
`
`5,320;102
`
`FIG. 12(a)
`
`HUMAN FEMORAL CARTILAGE INTENSITY
`PATIENT NAME: 1
`
`o AVG MED FEM ITNS
`oAVG LAT FEM INTNS
`
`1048
`1046
`
`1044
`1042
`
`1040
`
`1038
`
`1036
`
`INTENSITY
`
`1034 "t--r-'"T"'""""'"-r--r--r-"""'"!--".....,.-...-,.---r-......,
`-4
`-3
`-2 -1
`0 1 2 3 4
`LOCATIONS
`
`FIG. 12(b)
`
`HUMAN TIBIAL CARTILAGE INTENSITY
`PATIENT NAME: 1
`
`INTENSITY
`
`~ AVG MED TIB INTNS
`<> AVG LATTIB INTNS
`
`1044
`1043.5
`1043
`1042.5
`1042
`1041.5
`1041
`1040.5
`1040
`1039.5
`1039
`1038.5 +-...-~.......-r-......,...., ..............,...........,........,..........,.........,
`## -1 -.8 -.5 -.2 0 .25.5.75 11.2
`LOCATIONS
`
`-13-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 13 of 18
`
`5,320,102
`
`FIG. 13(a)
`
`HUMAN FEMORAL CARTILAGE INTENSITY
`PATIENT NAME: 2
`
`oAVG MED FEM ITNS
`oAVG LAT FEM INTNS
`
`1042
`1041
`1040
`1039
`1038
`INTENSITY 1037
`1036
`1035
`1034
`1033
`1032+---'.-.r---.--r---r----T""......-+--r---r---.-......--..---.-....,........,
`-1
`0
`1
`2
`3
`4
`-4
`-3
`-2
`LOCATION
`
`FIG. 13(b)
`
`HUMAN TIBIAL CARTILAGE INTENSITY
`PATIENT NAME: 2
`
`A AVG MED TIB INTNS
`o AVG LAT TIB INTNS
`
`1042
`1041
`1040
`1039
`1038
`INTENSITY 1037
`1036
`1035
`1034
`1033
`1032 +---.--.---.-..........--r-......-+--r-,--.--,.---..-,-....,........,
`-2
`-1
`0
`1
`2
`3
`4
`-4
`-3
`LOCATION
`
`-14-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 14 of 18
`
`5,320,102
`
`FIG. 14(a)
`
`HUMAN FEMORAL CARTILAGE INTENSITY
`PATIENT NAME: 3
`
`oAVG MED FEM ITNS
`oAVG LAT FEM INTNS
`
`-6
`
`4
`
`0
`~
`LOCATION
`
`2 468
`
`FIG. 14(b)
`
`HUMAN TIBIAL CARTILAGE INTENSITY
`PATIENT NAME: 3
`
`0 AVG MED TIB INTNS
`oAVG LAT TIB INTNS
`
`1048
`
`1046
`
`1044
`
`INTENSITY 1042
`
`1040
`
`1038
`
`1036
`-8
`
`1050
`
`1048
`
`1046
`
`1044
`INTENSITY 1042
`
`1040
`
`1038
`1036+---.~~-.----.--+-------..---......--,.--.---r-............,
`-4
`-3
`-2
`-1
`0
`1
`2
`3
`4
`LOCATION
`
`-15-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 15 of 18
`
`5,320;102
`
`FIG. 15(a)
`
`HUMAN FEMORAL CARTILAGE INTENSITY
`PATIENT NAME: 4
`
`o AVG MED FEM IT .
`oAVG LAT FEM INT ..
`
`1048
`
`1046
`
`1044
`
`1042
`INTENSITY 1040
`
`1038
`
`1036
`
`1034+-..--,---.--...........-----+--.-----r---..,.---.----,
`-8
`-6
`-4
`-2
`0
`2
`4
`6
`8
`LOCATION
`
`FIG. 15(b)
`
`HUMAN TIBIAL CARTILAGE INTENSITY
`PATIENT NAME: 4
`
`~ AVG MED TIB INTS
`o AVG LATTIB INTNS
`
`1043
`
`1042
`
`1041
`
`INTENSITY 1040
`
`1039
`
`1038
`
`1037 -+---""T""'""""'""""T""'""""'""~..,...........-+-..,...........'T""'"'""""""'T""'"'""""""....--.--.
`-2.5 -2 -1.5 -1
`-.5 0 .5 1 1.5 2 2.5
`LOCATIONS
`
`-16-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 16 of 18
`
`5,320,102
`
`FIG. 16(a)
`
`HUMAN FEMORAL CARTILAGE INTENSITY
`PATIENT NAME: 5
`
`+AVG MED FEM ITNS
`)( AVG LAT FEM INTNS
`
`1056
`1054
`1052
`1050
`1048
`INTENSITY 1046
`1044
`1042
`1040
`1038
`1036+---+-~.........,-+-r-......-~---.----.
`-4
`-3
`-2
`-1
`0
`1
`2
`3
`4
`LOCATIONS
`
`FIG. 16(b)
`
`HUMAN TIBIAL CARTILAGE INTENSITY
`PATIENT NAME: 5
`
`1052
`1050
`1048
`1046
`
`0 AVG MED TIB INTNS
`0 AVG LATTIB INTNS
`
`INTENSITY 1044
`1042
`1040
`1038
`1036
`1034+---.--r--.......--.---..---.--+-....--.-.......--..........-.---........
`-4
`-3
`-2
`-1
`0
`1
`2
`3
`4
`LOCATION
`
`-17-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 17 of 18
`
`5,320,102
`
`FIG. 17(a)
`
`HUMAN FEMORAL CARTILAGE INTENSITY
`PATIENT NAME: 6
`
`o AVG MED FEM ITNS
`o AVG LAT FEM INTNS
`
`1210
`1200
`1190
`1180
`1170
`INTENSITY 1160
`1150
`1140
`1130
`1120
`1110+----,--.......-.r--r--'"'.........-..........,f---o--r-.......,...-.-......--.
`-1.25
`-1
`-.75
`-.5
`-.25
`0
`.25
`.5
`.75
`1 1.25
`LOCATION
`
`FIG. 17(b)
`
`. HUMAN MEDIAL CARTILAGE INTENSITY
`PATIENT NAME: 6
`
`fj, AVG MED TIS INTNS
`<> AVG LATTIS INTNS
`
`1240
`
`1220
`
`1200
`
`1180
`INTENSITY 1160
`
`1140
`
`1120
`
`1100+-o--,--.....--,.-'"T"'""""""'-,..-......--+--....-----r---r--.-....-----.
`-2.5
`-2
`-1.5
`-1
`-.5
`0
`.5
`1 1.5
`2 2.5
`LOCATION
`
`-18-
`
`

`

`u.s. Patent
`
`June 14, 1994
`
`Sheet 18 of 18
`
`5,320;102
`
`FIG. 18(a)
`
`HUMAN INTENSITY
`PATIENT NAME: 7
`o AVG MED TIS INTNS
`o AVG LATTIS INTNS
`
`1047
`1046
`1045
`1044
`1043
`1042
`INTENSITY 1041
`1040
`1039
`1038
`1037
`1036+--.-.....--..--.--...----.-+----r-'"'--r-........-...........----.
`-4
`-3
`-2
`-1
`0
`1
`2 3
`4
`5
`-5
`LOCATION
`
`FIG. 18(b)
`
`HUMAN INTENSITY
`PATIENT NAME: 7
`
`1052
`1050
`1048
`1046
`
`o AVG MED FEM ITNS
`o AVG LAT FEM INTNS
`
`INTENSITY 1044
`1042
`1040
`1038
`1036
`1034 +-.--r-""......--.-,-.....-r-...-+-~---.-~..........,--.---,
`-4
`-3
`-2
`-1
`0 1 2
`3 4 5
`-5
`LOCATION
`
`-19-
`
`

`

`1
`
`5,320,102
`
`METHOD FOR DIAGNOSING PROTEOGLYCAN
`DEFICIENCY IN CARTILAGE BASED ON
`MAGNETIC RESONANCE IMAGE (MRI)
`
`BACKGROUND OF THE INVENTION
`1. Field of the Invention
`This invention relates to the diagnosis of proteogly(cid:173)
`can deficiency in articular cartilage based on magnetic
`resonance images (MRI) of the articular cartilage, and
`more particularly, based on a quantified signal intensity
`of each pixel of the magnetic resonance image extend(cid:173)
`ing across a depth of the articular cartilage.
`2. Description of the Related Art
`In magnetic resonance imaging (MRI), which is
`widely used for diagnostic purposes in medicine, a large
`magnet supported by complicated electronics and com(cid:173)
`puters is employed for image acquisition. When a pa(cid:173)
`tient is made to lie in the magnet's magnetic field, indi(cid:173)
`vidual atoms within the patient's tissues and organs
`become aligned with the magnetic field. Radiofre(cid:173)
`quency pulses are then transmitted at a defined rate to
`the magnetized atoms in the area of interest to thereby
`elevate the energy levels in the atoms. During a pause
`period between pulses, the atoms relax and part of the 25
`energy gained becomes released. This process repre(cid:173)
`sents the unique phenomenon of magnetic resonance. A
`receiving antenna collects the released energy signals
`which are then subjected to image processing to con(cid:173)
`vert the energy signals into light dots (pixels) having an
`illumination level (e.g. a gray-scale) corresponding to
`an intensity of the energy signals. Thousands and thou(cid:173)
`sands of such light dots form an image. The thus ob(cid:173)
`tained MR image is displayed on a television monitor,
`printed as a hard copy image and/or stored on a mag(cid:173)
`netic tape or other recording medium.
`An MR image provides structural information, such
`as size and shape, regarding most tissues and organs in
`the body, and permits the visual detection of changes in
`such structural characteristics. For example, if a person 40
`develops a meniscal or anterior cruciate ligament tear as
`a result of a sports related injury, an MRI of the joint
`will reveal
`the presence of a discontinuity, swelling
`and/or shortening in the image of the torn tissue. The
`structural
`image information provided by the MRI 45
`helps to decide whether a patient requires surgical re(cid:173)
`pair or other remedial action.
`FIG. 1 is a parasagittal section view (lateral to mid(cid:173)
`line) of the human knee joint. Reference numeral 101
`denotes the femur; 102 denotes the articularis genus 50
`muscle; 103 denotes the quadriceps femoris tendon; 104
`denotes 104 denotes the suprapatellar fat body; 105
`denotes the suprapatellar synovial bursa; 106 denotes
`the patella; 107 denotes the subcutaneous prepatellar
`bursa: 108 denotes the articular cavity; 109 denotes the 55
`infrapatellar fat body: 110 denotes the patellar ligament;
`111 denotes the synovial membrane; 112 denotes the
`subcutaneous infra patellar bursa; 113 denotes the deep
`(subtendinous) infrapatellar bursa; 114 denotes the lat(cid:173)
`eral meniscus; 115 denotes the tuberosity of tibia; 116 60
`denotes the bursa under lateral head of gastrocnemius
`muscle; 117 denotes the synovial membrane; 118 de(cid:173)
`notes the articular cartilages; and 119 denotes the tibia.
`Articular cartilage covers the opposing femur and tibia
`bone ends in the human knee joint. The articular carti- 65
`lage, which is rich in extracellular matrix and poor in
`cellularity, has shock absorption and lubrication func(cid:173)
`tions based on its visco-elasticity which depends on the
`
`2
`high water content of its extracellular matrix. Normal
`human articular cartilage has a water content of
`73%-81 % (w /w) on a weight basis. Proteoglycans (PG)
`are the vital organic component required for the func-
`5 tions of articular cartilage. PG contains numerous sugar
`chains, namely glycosaminoglycans, which contain
`negatively charged groups such as carboxylates and
`sulfate groups. These water absorbing, i.e, hydrophilic,
`groups attract an excess of water hydrogen atoms and
`10 water carrying cations. The wide spread network of PG
`retains this water in the cartilage matrix. A decrease in
`PG causes changes in the amount and state of water
`contained therein, resulting eventually in cartilage dys(cid:173)
`function. Thus a PG depletion is indicative of cartilage
`15 degeneration and precedes such problems as osteoar(cid:173)
`thritis.
`Although the MRI is used to visually detect struc(cid:173)
`tural changes in the articular cartilage of post-trauma
`joints, currently no non-invasive diagnostic tool is avail-
`20 able to detect a biochemical change such as PG deple(cid:173)
`tion in the articular cartilage at very early stages of
`cartilage degradation. Detection of PG depletion in
`cartilage prior to a structural change taking place could
`be extremely beneficial because steps could be initiated
`to preserve the cartilage by therapeutic intervention.
`Once PG depletion has started, the surface of the carti-
`lage begins to break after 1-2 years (fibrillation). As a
`result of fibrillation, usually in about 5 years the carti-
`30 lage becomes thinned resulting in a narrowing of the
`joint space. Therefore any attempt to protect or pre(cid:173)
`serve the cartilage must be made before initiation of
`surface breaking.
`Present techniques employing MRI are not capable
`35 of detecting biochemical changes, particularly PG de(cid:173)
`that develop in articular cartilage following
`pletion,
`joint trauma several years in advance of structural dis(cid:173)
`organization referred to as osteoarthritis (OA).
`
`SUMMARY OF THE INVENTION
`An object of the present invention is to provide a
`non-invasive method employing quantitative techniques
`for diagnosing a proteoglycan deficiency in articular
`cartilage, preferably prior to the onset of fibrillation and
`structural disorganization.
`Another object of the invention is to provide a non(cid:173)
`invasive method for tracking the progression or remis(cid:173)
`sion of proteoglycan depletion in articular cartilage
`which would be useful, for example, in a cartilage pre(cid:173)
`serving drug development/screening program.
`Another object of the invention is to provide a non(cid:173)
`invasive method for diagnosing an arthritic joint.
`In achieving the above and other objects, the method
`of this invention includes quantifying a signal intensity
`of a magnetic resonance image of the cartilage and
`correlating the thus quantified signal intensity with at
`least one predetermined reference signal intensity indic(cid:173)
`ative of cartilage proteoglycan content, e.g. with an
`expected or normal peak signal intensity indicative of an
`expected or normal proteoglycan content, or with an
`expected or normal signal intensity pattern indicative of
`an expected or normal proteoglycan concentration
`across cartilage depth.
`intensity is quantified
`the MRI signal
`Generally,
`across the depth of the cartilage as a gray-scale illumi(cid:173)
`nation of pixels of the image. The signal intensity varia(cid:173)
`tion across the depth of the cartilage is correlated with
`an expected or normal bell-shaped variation. Also, a
`
`-20-
`
`

`

`5,320,102
`
`3
`peak signal intensity in a middle portion of the cartilage
`is compared with predetermined reference signal inten(cid:173)
`sities indicative of expected or normal PG content.
`Further, the signal intensity within a same pixel layer
`of the cartilage may be analyzed, and a comparison of 5
`signal intensity variation in the medial vs. lateral con(cid:173)
`dyles may be carried out.
`
`4
`cess following a single injection is known to be reversed
`through increased synthesis by the chondrocytes which
`returns the cartilage proteoglycan content toward nor-
`mality. In the inventors' experiments in connection with
`the above-mentioned previous research, the depletion
`of proteoglycan content appeared to correlate with a
`decrease, and the repletion with an increase,
`in the
`cartilage thickness which was measured using the MR
`image.
`Since changes in cartilage thickness coincided with
`comparable changes in the measurable signal intensity, a
`quantitative MRI technique was developed and applied
`to healthy human knees to investigate whether the MRI
`signal intensity is truly related to the proteoglycan con(cid:173)
`tent. The goal was to quantify the signal intensity of all
`pixels in a particular region of the knee cartilage using
`computer-based image analysis. This was done in order
`to assess whether the signal intensity varied across the
`cartilage depth, and whether such variations corre(cid:173)
`spond with the distribution of any of the three major
`biochemical constituents of cartilage, i.e. water, colla-
`gen and proteoglycans, known to be differentially dis(cid:173)
`tributed across the cartilage depth.
`Six healthy volunteers ranging in age from 20-40
`years old were studied. The right knee joint of each
`individual was scanned using a l.5T (Signa, Ge, Mil(cid:173)
`waukee, Wis.) magnet and a dedicated transmit/receive
`extremity coil with the subject
`laying in the supine
`position. Three series of images were obtained in 5 of
`the 6 subjects, two using spin echo (SE) pulse sequences
`with TR mseclTE msec=700/20 and 1000/20. The
`third series was obtained using gradient refocused echo
`(GRE) 3D volume acquisition (60/15) and a flip angle
`of 15°. This pulse sequence was chosen since it reduces
`35 the scanning time while improving cartilage image con(cid:173)
`trast. The sixth subject was studied using only the GRE
`pulse sequence as explained below but using three dif(cid:173)
`ferent flip angles to further evaluate the signal intensity
`variation in cartilage images.
`For the spin echo, a coronal localizer (500/20) was
`followed by two sagittal acquisitions using first 700/20
`and then 1000/20 pulse sequences. Either an 8 or 12 em
`FOV was used. Other parameters were as follows: slice
`thickness 4 mm and interslice gap 1.5 mm, matrix
`256X 256 and two averages (in plane resolution 300 and
`450 microns). For the GRE sequence, the same FOV
`and plane were used. Twenty-eight to sixty contiguous
`slice locations were obtained using either 1.5 m or 3 mm
`slice thicknesses (in plane resolution 300 and 450 mi(cid:173)
`crons). Other parameters were kept similar for all the 5
`subjects studied. Finally, to further determine the de-
`gree of T1 weighted dependence of the signal intensity,
`in the sixth subject the knee was imaged using only the
`GRE sequence, but at 15°,30°, and 50° flip angles.
`For each study, femoral and tibial cartilage images
`obtained in the sagittal plane were analyzed on a Sun
`1+workstation (Sun Microsystems, Saddle
`Spare
`Brook, N.J.) using modified image display computer
`software that permitted x16 magnification of the MR
`cartilage image, and allowed measurement of the signal
`intensity on a pixel-by-pixel basis both across and along
`the full depth of the cartilage image.
`More particularly, reference is made to the imaging
`process flow chart of FIG. 2. As noted previously, the
`conventional MRI apparatus converts received signal
`intensities into image data denoting an illumination level
`of each pixel of the image corresponding to a location
`within the MRI slice or plane. This image data is re-
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`The object and features of this invention will become 10
`clear in the detailed description that follows with refer(cid:173)
`ence to the accompanying drawings in which:
`FIG. 1 is a parasagittal section view (lateral to mid(cid:173)
`line) of the human knee joint;
`FIG. 2 is a process flow chart illustrating the image 15
`processing of a magnetic resonance image according to
`the present invention;
`FIG. 3 is a femoral cartilage MR image at x16 magni(cid:173)
`fication showing individual pixels in the cartilage depth;
`FIGS. 4(a) through 4(c) illustrate MR images of a 20
`healthy knee obtained using two spin echo sequences
`and a GRE pulse sequence, respectively;
`FIGS. 5(a) through 5(j) respectively illustrate the
`signal intensities across the cartilage depth seen in im(cid:173)
`ages of the medial and lateral
`tibial plateau for five 25
`healthy volunteers and the mean for all healthy volun(cid:173)
`teers;
`FIGS. 6(a) through 6(j) illustrate the signal intensities
`across cartilage depth seen in images of the medial and
`lateral femoral condyles for the five healthy volunteers 30
`and the mean for all healthy volunteers;
`FIG. 7 illustrates the effect of changing the TRITE
`of the MR image on signal intensities across cartilage
`depth seen on the images of the lateral femoral condyle
`of the right knee of one of the healthy volunteers;
`FIG. 8 illustrates the affect of changing the flip angle
`on signal intensities across cartilage depth seen on the
`image of the medial femoral condyle of the right knee of
`one of the healthy volunteers;
`FIG. 9 illustrates variations in the distribution of 4D
`collagen, water and proteoglycan content across the
`depth of normal human articular cartilage;
`FIG. 10 illustrates the signal
`intensity across the
`depth of human articular cartilage and the proteoglycan
`content across the depth of the same cartilage;
`FIG. 11 is a table comparing the peak intensity from
`femoral and tibial cartilage of seven rheumatoid arthri-
`tis patients and five healthy volunteers; and,
`FIGS. 12(a) and 12(b) through FIGS. 18(a) and 18(b)
`respectively illustrate the signal intensity curve across 50
`the femoral cartilage and the signal intensity across the
`tibial cartilage of the seven rheumatoid arthritis pa(cid:173)
`tients.
`
`45
`
`55
`
`DETAILED DESCRIPTION OF THE
`PREFERRED EMBODIMENTS
`Previous research of the present inventors has shown
`that
`the cartilage thickness measurable on magnetic
`resonance images of rabbit knee cartilage, obtained
`using spin echo pulse sequences with TR= 1000 msec 60
`and TE=30 msec, correlated well with changes in the
`proteoglycan content brought about by intraarticular
`injection of the plant-derived proteolytic enzyme pa(cid:173)
`pain. (See Paul et al., "Magnetic Resonance Imaging
`Reflects Cartilage Proteoglycan Degradation in the 65
`Rabbit Knee", Skeletal Radiol, 1991.)
`Such injection of papain is known to profoundly
`deplete the cartilage proteoglycan content, but the pro-
`
`-21-
`
`

`

`5,320,102
`
`SPIN ECHO 1000/20
`
`6
`by the resolution of each pixel (e.g. 300 microns) pro(cid:173)
`vided a measure of cartilage thickness.
`The measured signal intensities from images of medial
`and lateral compartments of femoral and tibial carti(cid:173)
`lages obtained in the subjects with the 3 pulse sequences
`noted above were compared.
`Typical mid-lateral sagittal images of the same region
`of the right knee obtained using the three pulse sequen(cid:173)
`ces are shown in FIGS. 4(0) through 4(c). More particu-
`10 larly, FIG. 4(a) is an MR image of a healthy knee ob(cid:173)
`tained using a spin echo sequence TR= 1000 msec and
`TE=20 msec, showing the meniscus, femoral and tibial
`cartilage in the sagittal plane. FIG. 4(b) is a similar view
`obtained using a spin echo sequence TR=7oo msec and
`TE=20 msec. FIG. 4(c) is a similar view obtained using
`a GRE pulse sequence TR=60 msec and TE= 15 msec.
`Compared with spin echo images, cartilage contrast
`was greater in the GRE images. Also, as can be seen
`from FIG. 3, a x16 magnified image showed differential
`contrast across but not within the pixel layers. Across
`the pixel layers the .observed contrast was maximal in
`the middle and minimal at the surface and deep edges of
`the cartilage.
`As noted previously, the signal intensity of each pixel
`in the region was measured to obtain the average pixel(cid:173)
`by-pixel variation both across the pixel layers, i.e, along
`the cartilage depth, and within each pixel layer from a
`10 pixel wide area measuring between 3 and 4.5 mm-,
`depending upon the FOV used. Signal intensity within
`a pixel did not vary with magnification.
`The results for images obtained using the spin echo
`1000/20 sequence are described first in detail. Those for
`the other two sequences are then compared.
`
`5
`ceived at the workstation (step 1) and displayed (step 2)
`in a conventional manner. Then, the operator chooses a
`window or area of interest (e.g. the femoral cartilage)
`within the display image. The chosen area is then mag(cid:173)
`nified (step 3) for display. Responsive to user inputs 5
`individual pixels are selected (e.g by defining a region of
`interest within the overall magnified display image by
`tracing a boundary with the assistance of a computer
`mouse device), and the gray-scale illumination repre(cid:173)
`sentative of the MR signal intensity of selected individ(cid:173)
`ual pixels is recorded (step 4). Then, as will be discussed
`in more detail below, the recorded signal intensities of
`the selected pixels are subjected to appropriate averag-
`ing and interpolation (step 5), for the purpose of plot(cid:173)
`ting the signal intensity (y-axis) for each pixel along the 15
`depth (x-axis) of the cartilage (step 6).
`In this manner, signal intensity was measured from
`the individual pixels of the femoral and tibial cartilage
`images. For example, FIG. 3 is a femoral cartilage
`image with xl6 magnification showing individual pixels 20
`in the cartilage depth. The white arrow in FIG. 3 indi(cid:173)
`cates a direction across the cartilage (i.e. a "depth direc(cid:173)
`tion"), and the black arrow indicates a direction along,
`i.e. within, each cartilage pixel layer. (The pulse se- 25
`quence for the image obtained in FIG. 3 was GRE with
`TR=60 msec and TE= 15 msec.)
`Medial and lateral compartments were identified
`from anatomical and MR landmarks. Two mid-medial
`and two midlateral slices in the SE sequence, and five 30
`mid-medial and five mid-lateral slices in 3D volume
`GRE acquired cartilage images were analyzed. The
`region of the posterior femoral and tibial articular carti(cid:173)
`lages located above and below the middle third of a line
`connecting the base and the apex of the hypointense 35
`triangular shaped posterior horn of the medial and lat-
`eral meniscus was always analyzed.
`In each sagittal
`1. Pattern of Signal Intensity Variation
`Within pixel layers. Signal intensity within each pixel
`slice, signal intensity was measured from contiguous
`layer was found to vary randomly. The median value
`pixels along both the antero-posterior and supero-
`40 for the differences between adjacent pairs of pixels
`inferior planes (X and Y axes, respectively).
`equalled 7.0 (mean of 9.1 and mode of 5.0).
`The data were processed to obtain the following:
`Across pixel layers. Across the pixel layers, the signal
`1. Within pixel layer signal intensity variation (i.e. in
`the plane of the black arrow of FIG. 3): The differences
`intensity varied by a significantly greater margin, by as
`much as 200.0. FIGS. 5(0) through 5(j) respectively
`between the signal intensity of 411 adjacent pairs of
`pixels were measured and statistically evaluated.
`45 illustrate the signal
`intensities across cartilage depth
`seen in images of the medial and lateral
`tibial plateau
`2. Across pixel layer signal intensity variation (i.e. in
`obtained using a spin echo 1000/20 pulse sequence for
`the plane of the white arrow of FIG. 3): To minimize
`the five subjects and the mean for all subjects (mean
`errors from susceptibility artifacts, pixels at the edges
`differences between the signal
`intensities in cartilage
`near the bone and surface of the cartilage regions exam-
`ined were excluded from the assessment. Secondly, 50 zones 2 and 6 failed to be statistically significant). FIGS.
`since both the femoral and tibial cartilages have a
`6(0) through 6(j) illustrate the same in images of the
`curved geometry, the pixels are not oriented in a linear
`medial and lateral femoral condyles obtained using the
`spin echo 1000/20 pulse sequence. The signal intensity
`plane. Therefore, to facilitate the averaging process it
`was necessary to choose analysis of pixel columns con-
`was maximal in pixel layers of the middle zone and
`taining either even or odd numbers of units. Odd nurn- 55 minimal at both the superficial and deep edges. This
`bers were selected, and columns containing even num-
`resulted in the presence across the cartilage depth of a
`ber of pixels were interpolated across the entire depth to
`bell-shaped signal variation curve, which was present in
`an odd number. Thirdly, the signal intensity of 10 pixels
`all tibial and all except one out of ten femoral cartilages
`present
`in the same row antero-posteriorly (black ar-
`examined.
`row, FIG. 3) was averaged to obtain the mean signal 60
`intensity value for each zone or pixel layer. Such mea(cid:173)
`surements provided the inter-zonal or across the pixel
`layer signal variation curve in the cartilage region ex(cid:173)
`amined. This procedure was repeated in all of the me-
`dial and lateral slices studied.
`3. Pixel layers and cartilage thickness. The number of
`pixel layers present across the depth of the cartilage was
`counted at each of the 10 sites. The average multiplied
`
`2. Peak Signal Intensity
`The highest signal intensity was invariably present in
`pixel layers of the middle zone of the cartilage. In t