A standard test method for measurement of RF-induced heating (ΔT) on or near passive implants during MRI is described in ASTM F2182-11a [1]. It requires demonstrating the worst-case maximal temperature increase (max(ΔT)) without prescribing a method of worst-case analysis. In some reports worst-case analysis is based on searching cases with the highest value of SAR integrated over 0.1 gram of tissue (SAR_0.1g) and estimating max(ΔT) by temperature measurement at the location of the maximum (max(SAR_0.1g)). Our numerical simulation goals in this study were: 1) to compare locations of max(ΔT) and max(SAR_0.1g); 2) to evaluate the dependence of max(ΔT) on RF-induced power deposition at the hot spot (P_total) defined in ISO/TS 10974 [2]; 3) to analyze how ΔT changes along the path normal to the implant end face for two heating time points.The implants investigated were insulated stainless steel wires of 300 mm in length, of 0.73 and 1.6 mm in diameter, respectively. At one end of the wire, the insulation with thickness of 0.5 mm was removed, at the other side the wire was capped by the same insulation. The wires were placed in the ASTM phantom, which was located inside a 64 MHz high pass 16 rung birdcage coil with a diameter of 704 mm and a length of 650 mm. The ASTM phantom was filled with tissue-simulating medium (TSM) (ε_r = 78 and σ = 0.47 S/m). All wires were oriented along the Z axis, and shifted by 150 mm in X direction from the coil iso-center. The abbreviation "d×i×l" denotes: wire diameter “d”, outer insulation diameter “i”, length of insulation removed “l”. To evaluate the influence of wire end face shape on max(ΔT), wires with a chamfered wire end were also simulated (0.73×1.46×3_ch and 0.73×1.46×10_ch). ΔT(x,y,z), ΔT_tip, and ΔT_0.5mm were calculated as the temperature at location with coordinates x, y, z, the maximum wire temperature, and the temperature at location 0.5 mm from the wire tip (150, 0, 150.5), minus the ambient temperature of 22°C, respectively. P_total integration volume enveloped an area where volume loss density (VLD) decayed more than 20 dB from the peak value. A path from the wire tip (150, 0, 150) to a point (150, 0, 152) was defined to evaluate ΔT spatial variation for two heating time points (t = 10 and t = 900 s). All results were scaled to incident tangential electrical field E_tan = 1 V_peak-to-peak/m at the wire center.The results in Table 1 shows that correlations existed neither between P_total and SAR_0.1g nor between P_total or SAR_0.1g and max(ΔT). The area with highest VLD was located in close proximity to the wire tip (Fig. 2). Despite similar VLD distribution in this area, for a wire with l = 10 mm, a larger integration volume resulted in a higher P_total. All integration volumes were significantly smaller than the volume of 0.1 gram of TSM. The location of max(SAR_0.1g) was far away from the location of max(VLD) and max(ΔT), especially at t = 10 s. Thermal properties of two major thermal sinks - implant wire and TSM – varied considerably. The latter and different contact surface areas between the implant wire and TSM resulted in: 1) significant difference in thermal time constants, for example, the wires 0.73×1.46×3 and 0.73×1.46×3_ch heated up faster than other wires investigated (Fig. 3); 2) varying locations of max(ΔT); 3) essentially smaller ΔT_tip compared to max(ΔT) (Fig. 4); 4) temporal change in shape of the thermal hot spot (Fig. 5); 5) less heating of implants with a larger wire diameter.To fulfill the ASTM F2182-11a setup requirements, the temperature probe should be placed with submillimeter precision at a location that cannot be predicted by an EM simulation alone. Further, in some cases it also is impossible to predict the location of max(ΔT) at a given point in time by either SAR or the temperature measurement at another point in time. It is a challenge to validate with small uncertainty P_total calculated using EM simulation by only measuring SAR or VLD value at some points in space, if the field probe sensor size is larger than one tenth of the wire diameter. We doubt that an acceptable uncertainty of P_total can be achieved using thermal, SAR or VLD measurement only.