Research ArticleCLINICAL INVESTIGATIONS

Prediction Model of Chemotherapy Response in Osteosarcoma by 18F-FDG PET and MRI

Gi Jeong Cheon, Min Suk Kim, Jun Ah Lee, Soo-Yong Lee, Wan Hyeong Cho, Won Seok Song, Jae-Soo Koh, Ji Young Yoo, Dong Hyun Oh, Duk Seop Shin and Dae-Geun Jeon
Journal of Nuclear Medicine September 2009, 50 (9) 1435-1440; DOI: https://doi.org/10.2967/jnumed.109.063602

Article Figures & Data

Figures

  • FIGURE 1. 
    FIGURE 1. 

    ROC curve analysis of response prediction. ROC curves of SUV1 (A), SUV2 (B), SCR (C), VCR (D), and MVCR (E) were plotted to predict histologic response. On basis of AUC, all parameters—except SUV1—predicted histologic response.

  • FIGURE 2. 
    FIGURE 2. 

    Decision tree for response prediction. Decision tree was devised to predict histologic response based on SUV2 and MVCR values. Predictive values of our model for good responders and poor responders were 97% (31/32) and 95% (36/38), respectively.

Tables

  • TABLE 1

    Patient Characteristics

    CharacteristicValue
    Age (n)
     ≤15 y41 (58.6%)
     >15 and ≤40 y25 (35.7%)
     >40 y4 (5.7%)
    Sex (n)
     Male48 (68.6%)
     Female22 (31.4%)
    AJCC stage (n)
     IIA24 (34.3%)
     IIB40 (57.1%)
     IV6 (8.6%)
    Tumor volume (cm3)
     Median149
     Range17–2,882
    Location (n)
     Distal femur35 (50.0%)
     Proximal tibia17 (24.3%)
     Proximal humerus6 (8.6%)
     Others12 (17.1%)
    Pattern on plain radiograph (n)
     Lytic17 (24.3%)
     Blastic26 (37.1%)
     Mixed27 (38.6%)
    Pattern on MRI (n)
     Concentric60 (85.7%)
     Longitudinal10 (14.3%)
    SUV1
     Median8.0
     Range2.4–47.5
    SUV2
     Median4.5
     Range1.5–16.6
    Time from first PET to initiation of chemotherapy
     Median6 d
     Range1–13 d
    Time from end of chemotherapy to second PET
     Median19 d
     Range16–22 d
    Time from second PET to surgery
     Median2 d
     Range1–13 d
    Pathologic subtype (n)
     Osteoblastic50 (71.4%)
     Chondroblastic13 (18.6%)
     Fibroblastic5 (7.1%)
     Other2 (2.9%)
    Type of surgery (n)
     Amputation3 (4.3%)
     Limb salvage67 (95.7%)
    Histologic response (n)
     Good33 (47.1%)
     Poor37 (52.9%)

    • AJCC = American Joint Committee on Cancer.

  • TABLE 2

    Predictive Values of SUV2, SCR, VCR, and MVCR in 70 Patients

    ParameterCutoff valuenGR/PRPPV (%)NPV (%)Accuracy (%)
    SUV2≤277/01005963
    ≤31915/4796569
    ≤42923/6797677
    ≤54031/9789384
    ≤64631/15679276
    ≤75233/196310073
    SCR≤0.387/1885861
    ≤0.41614/2886570
    ≤0.52621/5817376
    ≤0.63526/9748077
    ≤0.74730/17648771
    ≤0.85030/20608567
    VCR≤0.81610/6635759
    ≤1.02922/7767374
    ≤1.24628/18617967
    MVCR≤0.287/1885861
    ≤0.41916/3846771
    ≤0.63428/6828684
    ≤0.84732/15689677
    ≤1.05232/20629470

    • GR = good responder; PR = poor responder; PPV = positive predictive value; NPV = negative predictive value; SCR = SUV2/SUV1; VCR = tumor volume after chemotherapy/tumor volume before chemotherapy; MVCR = SCR × VCR.

  • TABLE 3

    Summary of Previous Studies

    P1P2
    Referencen*GR/PRPMeanMedianRangeMeanMedianRangeCutoff†PPVNPV
    Schulte et al. (7)2717/10TBR10.33.3–33.22.920.32–20.33P2/P1 ≤ 0.6 for GR (n = 19)89.5%100%
    Franzius et al. (13)119/2TBR4.42.3–13.61.70.9–11.9P2/P1 < 0.7 for GR (n = 9)100%100%
    Ye et al. (16)158/7TBR7.13.0–20.63.11.1–6.5P2/P1 < 0.46 for GR (n = 8)100%100%
    Hawkins et at. (14)185/13SUVmax8.22.5–24.13.31.6–12.8P2 < 2 for GR (n = 4)75%85.7%
    Huang et al. (15)102/8SUVmax8.21.4–13.64.41.7–9.6P2/P1 < 0.4 for GR (n = 2)100%100%
    Present study7033/37SUVmax8.02.4–47.54.51.5–16.6Algorithm for GR (n = 32)97%95%

    • ↵* Patients with high-grade osteosarcoma were included in this table.

    • ↵† n = number of patients who met cutoff criteria.

    • GR = good responder; PR = poor responder; P = parameter; P1 = parameter before chemotherapy; P2 = parameter after chemotherapy; PPV = positive predictive value; NPV = negative predictive value; TBR = tumor-to-background ratio.

Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Bookmark this article

Jump to section