Table 1 Potential imaging biomarkers in the IPFP from OA patients. (+) and (-) indicate positive or negative association, respectively. AUC: Area under curve. CSA: Cross-sectional area. FSE: Fast spin echo. IPFP [H]: High IPFP signal intensity alteration. IVIM-DWI: Intravoxel incoherent motion diffusion-weighted MR imaging. KOOS: Knee injury and osteoarthritis outcome score. MAVRIC: Multiacquisition variable-resonance image combination. MEDIC: Multi-echo data image combination. ML: Machine learning. OARSI: Osteoarthritis Research Society International. PD-w: Proton density-weighted. PFJ: Patellofemoral joint. rOA: Radiographic OA. SGE: Spoiled gradient echo. SPAIR: Spectral attenuated Inversion recovery. SPGR: Spoiled gradient recall. SWE: Ultrasound shear elastography. sOA: Symptomatic OA. TKA: Total knee arthroplasty. TSE: Turbo spin echo. T1-w: T1-weighted. T2FS: fat-suppressed T2 hyperintense regions T2-w: T2-weighted. WOMAC: Western Ontario and McMaster universities osteoarthritis Index. WORMS: Whole-organ MRI score. ¹H-MRS: Hydrogen proton magnetic resonance spectroscopy

1) Cowan et al. [28]

Radiographic sOA, KL ≥ 2 (n = 35)

Asymptomatic control (n = 11)

Fat-suppressed T2-w MEDIC

IPFP volume in symptomatic group

2) Cai et al. [29]

rOA, KL ≥ 2 (n = 174)

Fat-saturated T1-w 3-D SPGR

Fat-saturated T2-w 2D FSE

(+) IPFP volume and cartilage volume

(+) IPFP volume and BML and osteophytes

3) Fontanella et al. [30]

End-stage OA (n = 28)

Patients meniscal tear (n = 32)

ACLR (n = 29)

Fat-suppressed T1-w and T2-w

 IPFP volume, surface, depth, and tibial arch length in end-stage OA

IPFP volume in ACLR group

IPFP hypointense signal in end-stage OA and ACLR

4) Ruan et al. [31]

sOA, KL ≥ 2 (n = 149)

MMP-13 ≤ (n = 75)

MMP-13 > (n = 74)

Fat-saturated T1-w 3-D SPGR

Fat-saturated T2-w 2D FSE

(-) Serum MMP-13 with IPFP and cartilage volume

(+) Serum MMP-13 with KL grading, IPFP [H], cartilage defect, serum IL-8, IL-18, TNFα

5) Chuckpaiwong et al. [32]

OA, KL = 2–3 (n = 15)

Control Healthy (n = 15)

Fat-suppressed T1-w 3D

No-fat-suppressed T2-w 3D

No differences in IPFP volume

OA IPFP volume increased with age

6) He et al. [33]

Clinical OA (n = 53)

Control Healthy (n = 54)

21 vs 21 matched by age, BMI, gender

3D T1-w FSE

3D PD-w fat-suppressed FSE

No correlation between knee pain and IPFP volume or area

(+) IPFP signal and cartilage loss

(-) IPFP signal and total pain

7) Steidle-Kloc et al. [34]

rOA kl 2–3 (n = 46)

Fat-suppressed

No association between IPFP volume and knee pain

8) Tan et al. [36]

sOA, KL = 2–3, WOMAC ≥ 4 (n = 84)

Asymptomatic OA, KL = 2–3 (n = 43)

Control Healthy, KL = 0–1 (n = 30)

IVIM-DWI

 IPFP depth in OA groups

IPFP [H] in the sOA compared to the asymptomatic OA

9) Fontanella et al. [37]

Late OA, undergoing TKA (n = 12)

Moderate OA, outerbridge score 3–4 undergoing meniscectomy (n = 15)

Control meniscal tears outerbridge score 0 (n = 17)

Fat-suppressed T2-w

 IPFP depth, femoral, and tibial length

Hypointense signal in moderate and late OA

10) Liu et al. [38]

OA KL ≥ 2 (n = 68)

Control KL = 0–1 (n = 41)

PD-w-SPAIR:

T2-w TSE

T1-w TSE

IPFP maximum CSA and IPFP depth in OA group

( +) IPFP [H] with age, meniscal injury, cartilage injury, and bone marrow edema

11) Han et al. [39]

OA (n = 977)

Fat-saturated T1-w 3D

(+) IPFP maximum area and cartilage volume

(-) IPFP maximun area and rOA

12) Satake et al. [40]

OA KL ≥ 2 (n = 97)

Patients with PFJ OA

Presence anterior knee pain (n = 41)

Absence anterior knee pain (n = 56)

SWE and

MRI: Fat-saturated T1-w 3D

Fat-suppressed T2-w 2D

(+) IPFP stiffness with anterior knee pain and femorotibial osteoarthritis

(-) -IPFP size and femorotibial osteoarthritis

13) Wang et al. [41]

sOA KL ≥ 2 (n = 45)

Control (n = 45) KL = 0–1

Fat-suppressed T2-w

(+) IPFP [H] and sOA

14) Carotti et al., [42]

Symptomatic OA (n = 149)

Fat-suppressed T1-w and T2-w

(+) WOMAC knee pain and IPFP synovitis

15) Ruan et al. [43]

rOA KL ≥ 2 (n = 160)

IL-8 ≤ median (n = 81)

IL-8 > median (n = 79)

Fat saturated T1-w 3D

Fat saturated T2-w 2D

 +) Serum IL-8, IPFP [H], and serum bone and/or cartilage biomarkers

16) Wang et al. [44]

rOA KL ≥ 2 (n = 170)

IL-17 ≤ median (n = 85)

IL-17 > median (n = 79)

Fat saturated T1-w SPGR

Fat saturated T2-w FSE

(+) (IPFP [H] with serum resistin and IL-17

17) Han et al. [45]

sOA (n = 200)

Fat saturated T1-w SGE

Fat-suppressed T2-w FSE

(+) Serum resistin with IPFP IPFP [H] and knee synovitis

18) Wu et al. [46]

sOA KL ≥ 2 (n = 146)

Ghrelin ≤ median (n = 74)

Ghrelin > median (n = 72)

Fat saturated T1-w 3D

Fat saturated T2-w 2D

(+) Ghrelin quartiles with IPFP IPFP [H], MMP3 and MMP13

19) Bian et al. [47]

sOA KL ≥ 2 (n = 137)

Citrate < median (n = 68)

Citrate ≥ median (n = 69)

Fat saturated T2-w

(-) Serum citrate with IPFP [H]

20) de Vries et al. [51]

OA undergoing TKA KL ≥ 2 (n = 22)

PFP (n = 35)

Healthy (n = 43)

T2 and DCE-MRI

73% OA patients showed T2FS-hyperintense IPFP regions

(+) IPFP T2FS-hyperintense regions with perfusion in OA patients

21) Han et al. [52]

OA (n = 874)

OARSI atlas

Fat-suppressed T1- or T2-w

(+) IPFP hypointense signals and rOA

(+) IPFP hypointense signals with cartilage defects and BMLs (longitudinal, 2.7 years)

22) Okita et al. [53]

OA KL = 1–4 (n = 15)

Healthy (n = 8)

T1- 3D MRI

IPFP contracture in OA

23) Chen et al. [54]

Advanced OA KL = 3–4 (n = 20)

Mild OA KL = 2 (n = 20)

No OA KL = 0–1 (n = 20)

T1-w PD SPAIR

3D six echo GRE

FF and T2* in end-stage OA

(-) FF and T2* and the BML, Hoffa-effusion synovitis, cartilage defect, total knee pain

24) Zhong et al. [55]

Advanced OA KL = 3–4 (n = 16)

Mild OA KL = 2 (n = 25)

Healthy KL < 1 (n = 23)

¹H-MRS

(-) FF and OA severity and Hoffa-synovitis

A weak inverse correlation with knee pain

Prognostic

25) Ruhdorfer et al. [56]

2 years

OA KL = 1–3 (n = 110)

Control no progression knees (n = 118)

Healthy (n = 88)

Intermediate-w fat-suppressed FSE

IPFP [H] in progressor OA knees

26) Harkey et al. [57]

2 years

Accelerated OA KL from 0–1 to 3–4 (n = 113)

No accelerated OA KL from 0–1 to 1–2 (n = 241)

Intermediate-w fat-suppressed TSE

Intermediate-w TSE, 3D dual-echo steady-state

Patients with increased IPFP [H] had a higher probability of developing end-stage OA

27) Davis et al. [58]

2 years

Accelerated OA (n = 125) KL from 0–1 to 3–4

Common OA (n = 125)

Control KL = 0–1 no changes in 4 years (n = 125)

Intermediate-w TSE fat-suppressed

IPFP [H] in end-stage OA compared to moderate OA, at 1 year before OA onset

28) Hill et al. [59]

0.5 years

rOA (n = 270)

Fat-suppressed T2-w

SE, PD

(+) Pain and IPFP synovitis

29) Roemer et al. [60]

5 years

Severe OA KL = 3–4 (n = 125)

No/mild OA KL ≤ 2 (n = 46)

Intermediate-w TSE 3D dual-echo

the 3D dual-echo at steady-state

Intermediate-w fat-saturated TSE

Hoffa synovitis was less frequent in No/mild rOA at baseline

Hoffa synovitis was similar between severe and No/mild rOA before TKA

30) Lu et al. [61]

sOA KL ≤ 3 (n = 100)

Fat-saturated

T2-w 3D SE

(+) IPFP sDev [H] and clustering factor [H], cartilage defect, bone marrow lesions and rOA

31) Wang et al. [62]

4 years

rOA KL ≥ 2 (n = 322)

Control No rOA in 4 years (n = 355)

Intermediate-w T2-w TSE

(+) IPFP Median [H], UQ [H], and the clustering factor [H] with incident rOA

(+) All measures with incident rOA 1 year prior OA detection

32) Cen et al. [63]

OA KL = 1–3 (n = 600)

Fat-saturated T2-w

(+) IPFP Mean [H] and Clustering factor [H] with radiographic and pain group

(+) IPFP [H] and radiographic group compared with pain group

33) Wang et al. [64]

5 years

OA underwent TKA after 5 years (n = 127)

Control no TKA after 5 years (n = 127)

Fat-saturated T2-w TSE

Association with TKA

Baseline: (+) Percentage (H)

1 year before TKA: (+) sDev [H], Percentage [H], and Clustering factor [H]

Before TKA: (+) all measurements

34) Han et al. [65]

- 2 years

sOA (n = 261)

Fat-suppressed T2-w FSE

Baseline: (+) sDev [H], UQ [H], and clustering factor [H] with tibiofemoral cartilage defects, and loss of tibial cartilage volume

35) Ruan et al. [66]

OA KL ≤ 3 (n = 255)

T2-w

(+) sDev [H], UQ [H], percentage [H], and clustering factor [H] with effusion-synovitis

36) Cen et al. [67]

OA KL = 1–3 (n = 600)

Fat-saturated T2-w

(+) Mean [H], sDev [H], Median [H], UQ [H], Percentage [H] and cartilage degradation (uC2C, uCTX-II)

bone turnover (uCTX-Iα and uNTX-I)

(+) Mean [H], Median [H] and UQ [H] with bone turnover (sCTX-I and uCTX-Iβ)

(+) Mean [H], Median [H]. and Percentage [H] with cartilage degradation (Coll2-1 NO2)

(+) sDev [H], Percentage [H] and inflammation (sHA)

No associations were found with Clustering factor [H]

37) Li et al. [69]

OA KL ≥ 2 4 years (n = 345)

Control no OA after 4 years (n = 345)

Voxel-based texture MRI

Diagnostic performance (AUC, 0.75)

38) Ye et al. [71]

Detectable OA KL ≥ 2 (n = 130)

No detectable OA KL < 2 (n = 34)

Radiomics

AUC of 0.78 in test datasets

(+) rad-scores and WORMS of cartilage, bone, meniscus, ligament, and synovium

39) Yu et al. [72]

OA KL ≥ 2 (n = 302)

Control KL = 0–1 (n = 302)

Radiomics

Diagnostic performance (AUCs, above 0.70)

40) Bonakdari et al. [73]

-

OA patients (n = 678)

High-BMI (n = 341)

Low-BMI (n = 337)

ML

Best models to predict IPFP volume: gender, age, and BMI, combined with

a) Total-cohort: adipsin/chemerin

b) High-BMI: chemerin/adiponectin HMW

c) Low-BMI: IL-8

Surgery outcome

41) Sacher et al. [74]

TKA (n = 28)

MAVRIC

T2 values in subjects with severe IPFP scarring

42) Cankaya et al. [76]

TKA

Total (n = 36)

Partial (n = 36)

Clinical and Isokinetic

Worse isokinetic performance

43) Gwyn et al. [77]

TKA

Total (n = 72)

Partial (n = 39)

Radiography

Patellar tendon lenght

44) Pinsornsak et al. [78]

TKA (n = 90)

Total (n = 45)

Partial (n = 45)

Clinical and sonographic (radiology)

No differences in patellar tendon shortening, and knee functionality

Anterior knee pain in resected group

45) İmren et al. [79]

5 years

TKA (n = 224)

Radiography

No differences in patellar tendon length

46) Michalak et al. [80]

0.5 years

TKA (n = 65)

Clinical and isokinetic

No differences in KOOS, functional outcomes, anterior knee pain, or patellar tendon length

47) Sellars et al. [81]

TKA (n = 111)

Radiography

No changes in patella tendon lenght