Concerns about Traces in Summing Circuit v2

[LoveLaika]

Previously, I built a summing circuit with op-amps, taking in 4 small signal voltages (~0.1V pulses) and outputting 3 sums/differences of those voltages. Now, I'm working on version 2 after learning from version 1. This time around, I still take in 4 inputs, but the sums are simplified to -(A+B), -(A+D), and -(A+B+C+D) using only one op-amp per sum. It's simple in nature, but I was wondering if I could get some feedback on my design so far, specifically regarding the traces from the inputs to the summing op-amps. I've attached my schematic so far along with my rough board layout.

For version 2, I removed my initial low pass filter at the input, and switched to current feedback op-amps for the bandwidth and speed. I've decided to use the LMH6703 op-amps. Aside from the reduction of the number of op-amps from version 1, I included linear regulators to provide power to my op-amps (ADP7182 for -5 volts, ADM7150 for +5 volts). I also removed some extraneous parts that I didn't need anymore, such as the voltage offset potentiometers. The input circuit protection components (protection from high voltages) are still present.

So, for my rough draft, I moved the over-voltage protection components to the bottom side of the board to save space. The power supplies haven't been routed yet, but right now, my concern is with the traces from the input to the op-amps. Running a test circuit in LTspice, my simulation shows that I should be able to achieve a very high bandwidth (at least 100 MHz) if I designed this correctly. That's what I'm hoping to achieve with this design.

With that in mind, using Saturn PCB (Bandwidth/Max Conductor Length), with a frequency of 100 MHz, I can go a distance of 428.2 mm before my traces can be considered a transmission line. At 500 MHz, that distance becomes 85 mm. Clearly, the shorter the distance, the better. With that in mind, I took into consideration the conductor impedance. The traces are referenced to a ground plane, so the shorter the distance my conductor is from the ground plane, the smaller my impedance becomes. A standard 2-layer board has 1.524 mm between the two copper planes while a 4-layer board has 0.1702 mm between the two top copper planes. Thus, I went with a 4-layer design, keeping my signal traces on the first two layers. I also went with larger trace widths (1.016 mm) to further reduce the impedance.

With the goal of achieving short distances in mind to reduce the impedances (and the max conductor length), you can see how I've laid out my circuit. The circuit protection components are on the bottom layer, away from other components, and the inputs are spaced closer to the op-amps, so they don't have to go through the circuit protection. That would add distance to the traces. I've also rearranged the op-amps and my inputs so that the traces from the inputs to the op-amps are kept short. The distances from the inputs to the op-amps are ~25 mm or less, with each trace being 1.016 mm in width. Since A was the one trace that went to all op-amps, I placed it on the second layer, reaching to its resistors via op-amps.

I still have a lot more to route, but this is the critical part for me, trying to achieve high frequency bandwidths by being careful with my routing. I was wondering if I could get some feedback of what I've done so far? Is there anything I can improve on? Are my assumptions correct regarding impedance, length, width, etc.?